Declining insulin requirement in the late first trimester of diabetic pregnancy.

OBJECTIVE: To investigate whether pregnancies complicated by type 1 diabetes are associated with a decrease in first-trimester insulin requirement. RESEARCH DESIGN AND METHODS: We examined the weekly insulin requirement (as units per kilogram per day) during the first trimester of pregnancy in diabetic women in the Diabetes in Early Pregnancy Study (DIEP) with accurate gestational dating, regular glucose monitoring, daily insulin-dose recording, and monthly glycohemoglobin measurements. RESULTS: In pregnancies that resulted in live-born full-term singleton infants, a significant 18% increase in mean weekly dosage was observed between weeks 3 and 7 (P = 0.000), followed by a significant 9% decline from week 7 through week 15 (P = 0.000). Further testing localized a significant change in insulin dose in the interval beginning weeks 7-8 and ending weeks 11-12 (P = 0.014). Within this interval, the maximum decrease was between weeks 9 and 10 (mean), 10 and 11 (median), and 8 and 9 (most frequent maximal decrease). To determine whether prior poor glucose control exaggerated these trends, we categorized the women based on their glycohemoglobin values: <2 SDs above the mean of a normal population (subgroup 1), 2-4 SDs (subgroup 2), and >4 SDs (subgroup 3) at baseline. Late first-trimester declines in dosage were statistically significant in subgroup 2 (P = 0.002) and subgroups 2 and 3 together (P = 0.003). Similarly, women with BMI >27.0 had a greater initial insulin rise and then fall compared with leaner women. CONCLUSIONS: Observations in the DIEP cohort disclose a mid-first-trimester decline in insulin requirement in type 1 diabetic pregnant women. Possible explanations include overinsulinization of previously poorly controlled diabetes, a transient decline in progesterone secretion during the late first-trimester luteo-placental shift in progesterone secretion, or other hormonal shifts. Clinicians should anticipate a clinically meaningful reduction in insulin requirement in the 5-week interval between weeks 7 and 12 of gestation.

The Diabetes in Early Pregnancy Study: beta-hydroxybutyrate levels in type 1 diabetic pregnancy compared with normal pregnancy. NICHD-Diabetes in Early Pregnancy Study Group (DIEP). National Institute of Child Health and Development.

OBJECTIVE: The objective was to assess relationships between beta-hydroxybutyrate (beta-OHB) level and pregnancy outcome in human pregnancy in light of the fact that high levels of beta-OHB cause malformations and growth retardation in in vitro studies. RESEARCH DESIGN AND METHODS: We analyzed beta-OHB in prospectively collected specimens from the National Institute of Child Health and Human Development-Diabetes in Early Pregnancy Study, in gestational weeks 6-12 in diabetic (n = 204-239) and nondiabetic (n = 316-332) pregnant women. RESULTS: Levels of beta-OHB in diabetic women were 2.5-fold higher than in nondiabetic pregnant women at 6 weeks' gestation and declined to 1.6-fold above nondiabetic women by 12 weeks' gestation (P < 0.0001 at all times). beta-OHB was positively correlated with glucose levels (P < 0.0001) in diabetic mothers, probably reflecting degree of diabetic control. beta-OHB correlated inversely with glucose (P < 0.0003) (gestational week 6 only) in nondiabetic mothers, possibly reflecting caloric intake. beta-OHB tended to be lower (not higher) in diabetic and nondiabetic mothers with malformed infants or pregnancy losses, but the difference was not statistically significant. beta-OHB in diabetic mothers at 8, 10, and 12 weeks correlated inversely with birth weight (P = 0.004-0.02), even after adjusting for maternal glucose levels. beta-OHB levels were also generally lower in diabetic mothers of macrosomic infants, and week 12 ultrasound crown-rump measurements were inversely related to beta-OHB levels. CONCLUSIONS: The lst trimester beta-OHB is significantly higher in diabetic than nondiabetic pregnant women. In both groups, beta-OHB tended to be lower, not higher, in mothers who had a malformed infant or pregnancy loss. beta-OHB was inversely related to crown-rump length and birth weight. The modest beta-OHB elevation in the 1st trimester of reasonably well-controlled diabetic pregnancy is not associated with malformations, probably because beta-OHB levels causing malformations in embryo culture models are 20- to 40-fold higher. The mechanism of the beta-OHB association with impaired fetal growth is unknown.

Methylenetetrahydrofolate reductase thermolabile variant and oral clefts.

Folic acid can prevent neural tube defects; in some cases the mechanism is probably a correction of a metabolic defect caused by thermolabile methylenetetrahydrofolate reductase (MTHFR) found in increased frequency in cases. It is less clear whether folic acid can prevent oral clefts, in part because it is not known whether thermolabile MTHFR is more common in those with oral clefts. This study examined the prevalence of the mutation (677 C-->T) that causes thermolabile MTHFR in subjects with oral clefts from a national Irish support group, and an anonymous control group randomly selected from a neonatal screening program covering all births in Ireland. Eighty-three of 848 control subjects were homozygous (TT) thermolabile MTHFR (9.8%). This defect was almost three times as common in the 27 subjects (25.9%) with isolated cleft palate (odds ratio 3.23, 95% confidence interval 1.32 -7.86, P = 0. 02) and somewhat more common in the 66 subjects with cleft lip with or without cleft palate (15.2%, odds ratio 1.65, 95% confidence interval 0.81-3.35, P = 0.20). When the two groups with different etiologies were combined, the overall odds ratio was 2.06 (95% confidence interval 1.16-3.66, P = 0.02). In the Irish population homozygosity for the common folate-related polymorphism associated with thermolabile MTHFR is significantly more frequent in those with isolated cleft palate, and could be etiologically important. Am. J. Med. Genet. 86:71-74, 1999. Published 1999 Wiley-Liss, Inc.

Prospective study showing that antisperm antibodies are not associated with pregnancy losses.

OBJECTIVE: To obtain prospective data on the relationship between presence of antisperm antibodies in maternal sera and first trimester pregnancy losses. DESIGN: First trimester sera obtained from diabetic and nondiabetic women recruited within 21 days of conception were analyzed using the indirect immunobead test for immunoglobulin (Ig)G, IgA, and IgM antisperm antibodies. Regional binding also was considered: sperm head, midpiece, and sperm tail. Results were correlated with pregnancy outcome. SETTING: Five university centers. PATIENTS: One hundred eleven women who experienced pregnancy loss (55 diabetic; 56 nondiabetic) were matched 2:1 with 104 diabetic and 116 nondiabetic women (controls) who subsequently had a normal liveborn infant. INTERVENTION: None. MAIN OUTCOME MEASURE: Pregnancy outcome (spontaneous abortion, liveborn) correlated with presence or absence of antisperm antibodies. RESULTS: Analyzing samples without knowledge of clinical status, no differences were observed with respect to IgG, IgA, and IgM binding when a positive test was defined as 50% of sperm showing antibody binding. Likewise, no association was found for IgG and IgM antisperm antibodies at 20% binding. The only positive finding was observed for IgA antisperm antibodies at 20% binding (Fisher's Exact test). This one positive finding merely could reflect multiple comparisons. No significant differences between groups were observed when analysis was stratified according to location of antibody binding (head, midpiece, tail tip, entire sperm). When the sample was stratified into those having and not having a prior loss, a relationship between antisperm antibodies and pregnancy loss still was not evident. CONCLUSION: Further work is necessary to determine whether IgA antisperm antibodies truly are associated with pregnancy loss or whether antisperm antibodies play any role in repetitive aborters. Findings in this study suggest that antisperm antibodies do not play a major role in pregnancy loss.

Lack of association between antiphospholipid antibodies and first-trimester spontaneous abortion: prospective study of pregnancies detected within 21 days of conception.

OBJECTIVE: To determine the role of antiphospholipid antibodies and anticardiolipin antibodies in first-trimester losses, addressing experimental pitfalls that preclude excluding the possibility that these antibodies reflect merely the selection bias of studying couples only after they have already experienced losses. DESIGN: Given that retrospective studies cannot exclude the possibility that such antibodies arise as a result of the fetal death, blood samples were obtained either before pregnancy or very early in pregnancy. Sera were obtained within 21 days of conception. SETTING: Multicenter university-based hospitals (National Institute of Child Health and Human Development collaborative study). PATIENT(S): Subjects for the current study were 93 women who later experienced pregnancy loss (48 diabetic; 45 nondiabetic), matched 2:1 with 190 controls (93 diabetic and 97 nondiabetic) who subsequently had normal live-born offspring. INTERVENTION(S): Sera from these 283 women were analyzed for antiphospholipid antibodies by enzyme immunoassay. In 260 of the 283 women (87 with pregnancy losses; 173 with live-born infants), sera were also available to perform assays for anticardiolipin antibodies by enzyme immunoassay. MAIN OUTCOME MEASURE(S): Pregnancy losses. RESULT(S): No association was observed between pregnancy loss and the presence of antiphospholipid antibodies or anticardiolipin antibodies. Levels of antiphospholipid antibodies were 6-19 PL/mL in 62.4% of the pregnancies that ended in losses and > or = 20 PL/mL in 5.4%; among pregnancies resulting in live-born infants, the percentages were 56.8% and 6.8%, respectively. Of the pregnancies that ended in a loss, 5.7% had anticardiolipin antibodies > or = 16 GPL/mL, compared with 5.2% of those ending in a live birth. CONCLUSION(S): This prospective study suggests that anticardiolipin antibodies and antiphospholipid antibodies are not associated with an increased risk for first-trimester pregnancy loss.

Periconceptional vitamin supplementation to prevent neural tube defects: how can we do it?

The discovery that folic acid can reduce neural tube defect rates offers a great opportunity for primary prevention. Unfortunately, women must receive the folic acid before or immediately after conception, before many know that they are pregnant. Thus, we are faced with a difficult choice: (1) ask all women at risk of getting pregnant to take supplements, or (2) fortify the food supply to ensure that all women at risk receive additional folic acid. Neither approach is ideal. Many women will not take vitamin supplements. Fortification at sufficiently high levels to provide all women with 400 micrograms of folic acid will expose other segments of the population to unacceptably high levels. Because many women of child-bearing age are unaware of the benefits of folic acid, a vigorous education campaign should begin immediately to encourage women at risk to take supplements. Adding 70 micrograms of folic acid per 100 g of grain could be justified easily because this amount is removed from grain in processing. If it is technically feasible, adding up to 140 micrograms is likely to be safe, and could prevent more NTDs. A major educational campaign and modest fortification of grain with folic acid may be the best practical solution.

Folic acid to prevent neural tube defects: scientific advances and public health issues.

Folic acid prevents some neural tube defects by correcting abnormal homocysteine metabolism. A defective gene coding for a homocysteine-related enzyme has been shown in people with neural tube defects. An extensive educational campaign will be required to alert women to the need to take 0.4 mg/day of folic acid during their childbearing years.

Vitamins, folic acid and neural tube defects: comments on investigations in the United States.

No clear answer concerning whether multivitamin/folate supplementation prevents neural tube defects (NTDs) is provided by three studies in the United States. All these studies are occurrence in nature, no recurrence studies having been conducted. The Atlanta Birth Defects Study is subject to pronounced memory and recall biases, the length between event and interview being as long as 16 years. In a second study (Boston University), objections can be raised to certain aspects of the experimental design, and the claim that 22 per cent of women started vitamins sufficiently early after pregnancy diagnosis to influence NTD formation is suspicious. Our NICHD case control study of 541 women in California and Illinois revealed no evidence for multivitamins or folic acid preventing NTDs. U.S. public policy-makers face difficulties in applying results of recurrence or occurrence studies in high-risk areas to low-risk areas in the U.S.

Genetic heterogeneity in neural tube defects.

In 1985-1987, the authors attempted to ascertain all cases of confirmed neural tube defects (NTD) in California and Illinois, not only among live-born infants (postnatal) but also cases ascertained during pregnancy (prenatal). Mothers of both prenatal and postnatal NTD cases were interviewed within 5 months. Among postnatal NTD cases, 14.9% (45/303) had anomalies not ordinarily associated with NTD. The frequency of non-NTD related anomalies was 9.4% (5/53) in anencephaly, 0/3 in craniorachischisis, 22.9% (8/35) in encephalocele, 14.5% (27/186) in spina bifida, 20% (1/5) in multiple NTD cases and 19% (4/21) in other NTDs. However, relatively few postnatal NTD cases had known multiple malformation patterns; Meckel-Gruber syndrome was the most common, with 2 postnatal cases, and 3 additional prenatal cases. Maternal age, paternal age and birth order in postnatal cases were 26.7 +/- 5.4 SD, 28.9 +/- 5.8 and 2.8 +/- 1.8, respectively. These characteristics were similar in prenatal NTD cases (27.9 +/- 6.0, 30.1 +/- 6.3, 2.5 +/- 1.5, respectively). We also found no differences in parental ages among different types of NTD. Frequency of prior spontaneous abortion differed neither between postnatal NTD (9.3%) and postnatal controls (8.1%), nor between prenatal NTD (10.7%) and prenatal control (8.7%). Loss rates in the pregnancy immediately prior to the index NTD cases were not significantly higher than in control subjects. The high frequency of non-NTD associated malformations (14.9%) indicates the caution must be exercised before assuming that a given NTD case is polygenic-multifactorial in etiology, especially cases of encephalocele.

Vitamin A and birth defects.

OBJECTIVE: Our objective was to determine whether moderate doses of vitamin A are teratogenic. STUDY DESIGN: This was a geographically based case-control study. Women whose pregnancies produced offspring with neural tube defects (n = 548) or major malformations other than neural tube defects (n = 387) and normal control subjects (n = 573) were interviewed to determine periconceptional vitamin A supplement exposure levels. RESULTS: The proportion of women consuming doses of vitamin A between 8000 and 25,000 IU was no greater in the major malformations group or the group with neural tube defects than in the normal control group. For exposure from supplements and fortified cereals combined, women consuming >8000 and >10,000 IU daily had odds ratios for major malformations of 0.79 (95% confidence interval 0.40 to 1.53) and 0.73 (95% confidence interval 0.27 to 1.96), respectively, compared with women consuming <5000 IU. The results for neural tube defects were similar. For cranial neural crest defects the odds ratios were 0.76 (0.22 to 2.56) and 1.09 (0.24 to 4.98) for exposure to >8000 and >10,000 IU, respectively, versus exposure to <5000 IU. CONCLUSIONS: This study found no association between periconceptional vitamin A exposure at doses >8000 IU or >10,000 IU per day and malformations in general, cranial neural crest defects, or neural tube defects. If vitamin A is a teratogen, the minimum teratogenic dose appears to be well above the level consumed by most women during organogenesis.

Risk of neural tube defects in relation to maternal fertility and fertility drug use.

In a case-control study to investigate whether women who use drugs to induce ovulation are at increased risk of conception of a child with a neural tube defect, 571 women who had a fetus or child with a neural tube defect, 546 women who had a fetus or child with other abnormalities, and 573 women who had an apparently normal fetus or child were questioned about infertility, fertility drug use, and related obstetric problems. The rate of maternal fertility drug use around the time of conception was not significantly higher for neural tube defects than for other abnormalities (odds ratio 1.28; 95% confidence interval 0.39, 4.51) or no abnormalities (odds ratio 0.80; 95% Cl 0.27, 2.27). Fertility drug use at any time was not significantly more frequent for neural tube defects than for other abnormalities (odds ratio 1.37; 95% Cl 0.70, 2.74) or no abnormalities (odds ratio 1.05; 95% Cl 0.56, 1.98).

Thermolabile variant of 5,10-methylenetetrahydrofolate reductase associated with low red-cell folates: implications for folate intake recommendations.

BACKGROUND: The dietary reference values for folate, as for other nutrients, are targeted to the general and supposedly normal population, not people with special needs, such as those with genetic or metabolic abnormalities or diseases. However, 5-15% of general populations are homozygous for a thermolabile variant of 5,10-methylenetetrahydrofolate reductase (C677T) which causes mild hyperhomocysteinaemia and is positively associated with the development of vascular disease and the risk of neural-tube defects. If tissue-folate status is compromised in large sectors of the population by this or other genetic variants, the present dietary reference values may need to be changed. METHODS: We identified the C677T genotype and measured red-cell folate concentrations in two groups of healthy women (pregnant, 242, not pregnant, 318). We then analysed the effect of genotype on red-cell folates, which are a reliable marker for tissue folate stores. FINDINGS: In the pregnant group there were 20 TT homozygotes, 114 wild-type CC homozygotes, and 108 CT heterozygotes. In the non-pregnant group, the numbers were 41, 148, and 129. In both pregnant and non-pregnant groups, red-cell folate was significantly lower among TT homozygous than CC homozygous women (mean 252 [95% CI 202-317] vs 347 [321-372] micrograms/L, p = 0.002 for pregnant women; 284 [250-327] vs 347 [342-372] micrograms/L, p = 0.01 for non-pregnant women). Plasma folate was also significantly lower in TT homozygous than in CC homozygous women in the pregnant group (p = 0.009) but not in the non-pregnant group. INTERPRETATION: These results suggest that a substantial minority of people in general populations may have increased folate needs. Future studies may show the presence of other common genetic variants that interact with particular nutrients and place doubts on the validity of assuming "normality" for nutrient requirements in any general population.

Are obese women at higher risk for producing malformed offspring?

OBJECTIVE: Our purpose was to determine whether obese women and underweight women have an increased risk of birth defects in their offspring. STUDY DESIGN: A geographically based case-control study of women living in California and Illinois was performed. There were 499 mothers of offspring with neural tube defects, 337 mothers of offspring with other major birth defects, and 534 mothers of offspring without birth defects who participated. RESULTS: Compared with women of normal weight, women who were extremely obese before pregnancy (body mass index > or = 31 kg/m2) showed a significantly increased risk of having an infant with a neural tube defect (odds ratio 1.8, 95% confidence interval 1.1 to 3.0), especially spina bifida (odds ratio 2.6, 95% confidence interval 1.5 to 4.5), after adjustments for age, race, education, and family income. Obese women also had significantly increased risks (p < 0.05) of having an infant with other defects of the central nervous system, great vessel defects, ventral wall defects, or other intestinal defects. CONCLUSION: Our data suggest that offspring of obese women (but not underweight women) are at an increased risk of neural tube defects and several other malformations. If these findings are confirmed, further research will be necessary before it can be concluded that weight reduction before pregnancy will lower the risk of birth defects among obese women. Until then, obese women can address their risk of birth defects with the same measures that are recommended for all women, such as adequate daily intake of folic acid and alpha-fetoprotein screening to identify malformed fetuses.

Homocysteine metabolism in pregnancies complicated by neural-tube defects.

Folic acid taken around the time of conception can prevent many neural-tube defects. Women with low-normal vitamin B12 values may also be at increased risk. We considered whether homocysteine metabolism via the enzyme methionine synthase, which requires both folate and B12, could be the critical defect in folate-related neural tube defects. Blood was obtained during pregnancies that produced 81 infants with neural-tube defects and 323 normal children. Samples were assayed for homocysteine, methylmalonic acid, plasma folate, red-cell folate, and B12. Mothers of children with neural-tube defects had significantly higher homocysteine values (8.62 [SD 2.8] mumol/L) than did B12-matched controls (7.96 [2.5] mumol/L, p = 0.03). The difference was significant (p = 0.004) in the lower half of the B12 distribution after adjusting for plasma folate. Our study shows that an abnormality in homocysteine metabolism, apparently related to methionine synthase, is present in many women who give birth to children with neural-tube defects. Overcoming this abnormality is likely to be the mechanism by which folic acid prevents neural-tube defects. These findings suggest that the most effective periconceptional prophylaxis to prevent neural-tube defects may require B12 as well as folic acid.

Homocysteine and neural tube defects.

It is now well established that folic acid, when taken periconceptionally, can prevent many neural tube defects. It is also becoming clear that folic acid does not work by correcting a nutritional deficiency in pregnant women. Rather, it appears that a metabolic defect is responsible for these neural tube defects and that this defect or defects can be corrected by a sufficiently large dose of folic acid. Our recent work demonstrates that homocysteine metabolism is likely to be the critical pathway affected by folic acid. We have demonstrated significantly higher homocysteine levels in women carrying affected fetuses than in control women. These findings indicate that one of the enzymes responsible for homocysteine metabolism is likely to be abnormal in affected pregnancies. Animal studies suggest that the conversion of homocysteine to methionine could be the critical step. Rat embryos in culture require methionine for neural tube closure. Methionine synthase, cystathionine synthase, and 5,10 methylene tetrahydrofolate reductase are all important in the metabolism of homocysteine in humans. If methionine synthase is the critical enzyme, it would raise the interesting public health issue that vitamin B-12 might be able to stimulate the abnormal enzyme as folic acid does. Adding vitamin B-12 might make it possible to reduce the dose of folic acid required in fortified food, thus allaying concerns about overexposure to folic acid.

The Diabetes in Early Pregnancy Study: changes in cholesterol, triglycerides, body weight, and blood pressure. The National Institute of Child Health and Human Development--the Diabetes in Early Pregnancy Study.

This study examined changes in cholesterol, triglycerides, body weight, and blood pressure during pregnancy in 312 diabetic and 356 control women recruited within 21 days after conception. Cholesterol values rose in both groups but were significantly lower in diabetic women at each time point (166 vs 178 mg/dl at week 12, p = 0.0004). Triglyceride values also rose in both groups. Triglyceride levels did not differ between groups up to week 8 of gestation, but by weeks 10 to 12 they were significantly lower in diabetic women than in controls (75 vs 89 mg/dl at week 12, p = 0.0004). Although they were no heavier at entry, diabetic women gained significantly more weight between weeks 6 and 8 (p less than 0.001), resulting in a mean difference between groups of 1 kg. Systolic blood pressure increased steadily and significantly in the diabetic but not the control women (115.8 +/- 16.2 SD vs 109.3 +/- 11.8 mm Hg, p = 0.0006 at term). Diastolic blood pressure was higher in diabetic women on entry (70.7 vs 67.3 mm Hg, p = 0.0006) and throughout gestation. Significant correlations were found in the diabetic group between maternal blood pressure and lipids and infant birth weight. These newly found differences in cholesterol and triglyceride levels, weight gain, and blood pressure between type I diabetic and control women during gestation may have long-term cardiovascular implications.

The absence of a relation between the periconceptional use of vitamins and neural-tube defects. National Institute of Child Health and Human Development Neural Tube Defects Study Group.

Whether taking multivitamins or folate around the time of conception can reduce a woman's risk of having a child with a neural-tube defect is controversial. To investigate this question, we examined the periconceptional use of vitamin supplements by women who had a conceptus with a neural-tube defect (n = 571), women who had had a stillbirth or a conceptus with another malformation (n = 546), and women who had had a normal conceptus (n = 573). Women with conceptuses with neural-tube defects were identified either prenatally or postnatally and were matched to control mothers for gestational age. To minimize recall bias, we interviewed nearly all the women within five months of the diagnosis of a birth defect or the birth of the infant (mean, 84 days); information on vitamin use was obtained by an interviewer who was unaware of the outcome of pregnancy. The rate of periconceptional multivitamin use among the mothers of infants with neural-tube defects (15.8 percent) was not significantly different from the rate among mothers in either the abnormal or the normal control group (14.1 percent and 15.9 percent, respectively). After adjustment for potential confounding factors, the odds ratio for having an infant with a neural-tube defect among women classified as having had full supplementation with multivitamins was 0.95 as compared with the mothers of the abnormal infants (95 percent confidence interval, 0.78 to 1.14) and 1.00 as compared with the mothers of normal infants (95 percent confidence interval, 0.83 to 1.20). There were no differences among the groups in the use of folate supplements. The adjusted odds ratio for having an infant with a neural-tube defect among those receiving the recommended daily allowance of folate was 0.97 as compared with the mothers of abnormal infants (95 percent confidence interval, 0.79 to 1.18) and 0.98 as compared with the mothers of normal infants (95 percent confidence interval, 0.80 to 1.20). We conclude that the periconceptional use of multivitamins or folate-containing supplements by American women does not decrease the risk of having an infant with a neural-tube defect.

Moderate caffeine use and the risk of spontaneous abortion and intrauterine growth retardation.

OBJECTIVE: To examine the relationship between caffeine consumption during pregnancy and the occurrence of spontaneous abortion and intrauterine growth retardation. DESIGN, SETTING, AND PATIENTS: A cohort of 431 women, enrolled in a multicenter study within 21 days of conception, was monitored throughout pregnancy to determine (1) caffeine exposure, (2) exposure to other risk factors, (3) fetal growth as assessed by ultrasonography, and (4) pregnancy outcome. OUTCOME MEASURES: Spontaneous abortion, intrauterine growth, birth weight, and head circumference. RESULTS: The mean (+/- SD) first-trimester caffeine consumption was not significantly higher in women who aborted (125.9 +/- 123.1 mg) than in women who delivered liveborn infants (111.6 +/- 107.0 mg) (P = 34). The adjusted odds ratio (OR) for spontaneous abortion was 1.15 (95% confidence interval [CI], 0.89 to 1.49). Early fetal growth, assessed by crown-rump length on ultrasonographic examination, was not affected by caffeine. Although the group consuming the most caffeine (> 300 mg/d) had a significantly higher proportion of babies with birth weights and head circumferences below the 10th percentile in the crude analysis, the association with caffeine was no longer significant when other risk factors (notably smoking) were taken into account. The adjusted ORs were 1.11 (95% CI, 0.88 to 1.40) for decreased birth weight and 1.09 (95% CI, 0.86 to 1.37) for smaller head circumference. CONCLUSIONS: Close monitoring of a cohort identified very soon after conception enabled us to identify all abortions after 21 days postconception, monitor intrauterine growth prospectively, and track caffeine use. Despite this intensive surveillance, we found no evidence that moderate caffeine use increased the risk of spontaneous abortion, intrauterine growth retardation, or microcephaly after accounting for other risk factors.