OBJECTIVE: Previous studies have reported an increase in twinning of as much as 40% associated with folic acid–containing supplements, and folic acid fortification of enriched cereal grains was authorized in 1996. The purpose of this study was to investigate whether twinning rates have increased since that time.
METHODS: We used United States birth and fetal death records to calculate twin gestation rates from 1990 through 2000. To eliminate the influence of fertility treatments, our analysis was limited to nulliparous women aged 16–19. We compared time trends in twin gestation rates before and after folic acid fortification in 1996.
RESULTS: A total of 25,065 twin and 3,362,245 singleton pregnancies were included. Twin gestation rates were stable from January 1990 through December 1996, at 7.2 per 1,000, and then began a steady increase, averaging 2.4% (95% confidence interval 0.1–4.2%, P = .006) per year, which continued through 2000 and reached 8.2 per 1,000. This translates to 2 additional twin pregnancies per 10,000 gestations per year. Twin rates continued to increase well beyond 1998, when the maximal fortification effect on folate status had been reached.
CONCLUSION: Although twin gestation rates in women not using fertility treatments increased after food fortification with folic acid, they rose by much less than the 40% rate previously reported; the observed pattern of increase in twin gestation rates is not consistent with a folic acid fortification effect.
LEVEL OF EVIDENCE: II-2
Although twin gestation rates increased between 1990 and 2000, the pattern of increase is not consistent with a folic acid fortification effect.
From the *Epidemiology Branch, †Biometry and Mathematical Statistics Branch, and ‡Computer Sciences Section, Division of Epidemiology, Statistics and Prevention Research, The National Institute of Child Health and Human Development, National Institutes of Health, Health and Human Services, Bethesda, Maryland.
Reprints are not available. Address correspondence to: James L. Mills, MD, MS, Chief, Pediatric Epidemiology Section, 6100 Executive Boulevard, Room 7B03, DESPR, NICHD, NIH, Bethesda, MD 20892; e-mail: firstname.lastname@example.org.
Received July 21, 2004. Received in revised form October 25, 2004. Accepted December 16, 2004.
Randomized trials have shown that periconceptional folic acid supplementation can decrease the occurrence of neural tube defects (NTDs) by as much as 70%.1,2 In response to this observation, the United States Public Health Service recommended that all women of reproductive age consume 400 μg of folic acid daily to reduce the incidence of NTDs.3 Unfortunately, most women did not follow the recommendation.4 To increase folic acid intake in women of reproductive age, the U.S. Food and Drug Administration mandated the fortification of enriched cereal grain products at 140 μg of folic acid per 100 g of grain.5 Voluntary fortification was started in March 1996; by January 1998, full fortification was mandatory. National data indicate that among women of reproductive age, serum folate levels have risen 171% since fortification.6 Ray7 showed that, in Canada, where the fortification plan is almost identical to that of the United States, serum folate levels rose quickly until late 1998, and then reached a plateau. Fortification has clearly brought population folate levels to protective range, as the incidence of NTDs has decreased by 19–54%.8–10
Several studies11–13 have reported an increase in twinning associated with use of multivitamin supplements containing folic acid. Because twin pregnancies are at greater risk for maternal morbidity and infant morbidity and mortality,14 an increase in the twinning rate caused by population-wide exposure to folic acid could have significant public health implications.15 In 1994, Czeizel et al11 reported a 40% increase in the number of twin births among Hungarian women randomized to treatment with periconceptional multivitamins that included 0.8 mg of folic acid, compared with women treated only with trace elements. Similar results were noted in observational studies in Sweden12 (odds ratio 1.45. 95% confidence interval [CI] 1.06–1.98) and the United States,13 although the increase in the U.S. study was not statistically significant. Other studies have shown no increase in twinning associated with food fortification in the United States.16,17
Because of the risks associated with twinning, a clinically significant folic acid-induced increase could stimulate a reconsideration of fortification policy. In this study, we used national natality and fetal death data to examine twin gestation rates before and after the introduction of folic acid fortification to the U.S. food supply to estimate 1) whether twin pregnancy rates increased after fortification and 2) if they had, whether the pattern of increase was consistent with a folic acid effect, ie, whether the twin rates increased as the maternal population's body folate stores increased, then reached a plateau as folate stores reached a plateau.
MATERIALS AND METHODS
Natality data from the National Vital Statistics System/National Center for Health Statistics for the years 1990–2000 were used to determine the number and plurality of all live births in the United States. These data sets are compiled from 100% of all registered birth certificates from the 50 states and the District of Columbia.18,19 Fetal deaths were identified from fetal death files compiled by National Vital Statistics System/National Center for Health Statistics for the years 1990–1994.20 From 1995 on, fetal death data have been included as part of perinatal mortality data files.21 The number and plurality of stillbirths occurring between 1990 and 2000 in all states and the District of Columbia were abstracted from these sources.18–21 Pregnancies ending in spontaneous abortion and elective termination were not included.
Reporting in natality and fetal death data sets is by infant, not by delivery. Triplet and higher-order births were excluded from this analysis (see below). A twin delivery is defined as delivery after 20 weeks of gestation of 2 clinically recognized fetuses, regardless of the vital status of either infant at birth. Each live and stillborn twin and singleton birth from 1990 to 2000 was categorized according to plurality, maternal age (15–44 years), race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, Asian), parity (0 or 1+), and month and year of birth. The number of twin deliveries for each year was calculated as the total number of first-born twin infants. The total number of deliveries was then calculated as the sum of the total number of singleton deliveries plus the calculated number of twin deliveries. Twin gestation rates were calculated as the number of live or stillborn twin deliveries/total number of deliveries for each age and racial group.
Data on fertility treatments, an important cause of multiple gestations, are not collected on all women. Therefore, to minimize the influence of assisted reproduction (ovulation induction and/or assisted reproductive technologies [ARTs]) on multiple gestations, our initial target population was nulliparous women aged 15–19 years because this group is unlikely to be receiving fertility treatment. Using data compiled by the Society for Assisted Reproductive Technology and the Centers for Disease Control, Reynolds and colleagues22 reported that 0.0% of infants born to women younger than the age of 20 years were conceived using ART between 1997 and 2000. The 1988 National Survey of Family Growth data showed that 8.4% of nulliparous women aged 15–24 years reported impaired fecundity.23 Data from the same survey indicated that of the total number of women aged 15–19 with self-reported impaired fecundity, 2.1% had ever obtained specialized infertility services, such as ovulation induction, intrauterine insemination, or in vitro fertilization.24 Therefore, we estimated that less than 0.2% of teenagers receive fertility treatments. Additionally, records with plurality of triplet or greater were excluded because these high-order births are usually (∼80%) the result of fertility treatments.25
Because the numbers of twin pregnancies among all women aged 15 years (n = 831) and among Asian women aged 15–19 years (n = 347) were relatively small, making the twin gestation rates imprecise, these 2 groups are not reported. Therefore, our final analysis group consisted of nulliparous, non-Hispanic white, non-Hispanic black, and Hispanic women aged 16–19 years.
Folic acid might increase twin pregnancy by increasing the number of twin conceptions. To examine this hypothesis, we assumed that infants born in December 1996 or after were likely to have been conceived on or after March 1996, the month that folic acid fortification was first authorized. We defined these births as exposed to fortification. We assumed that births before December 1996 were conceived before fortification was introduced, and defined these births as unexposed.
Changes in the twinning rate over the course of time were adjusted for maternal age (in years) and race/ethnicity, as well as their interaction, using a log linear (Poisson regression) model. This is a standard method for the analysis of rates in which the numerator (number of twin pregnancies in a given category) serves as the dependent variable, with adjustment for the denominator (total number of pregnancies in the category) as an offset.26 The regression model included different time trends for the 2 time periods (before and after December 1996). The full model then consisted of, for each combination of age and race, a broken line with separate slopes for the prefortification and postfortification periods.
The null hypothesis of interest was that the time trends in the 2 periods were identical (a single line rather than a broken line with 2 different slopes before and after December 1996), indicating no change in twin gestation rate associated with folic acid fortification. Statistical analysis was performed with SAS 8.2 software (SAS Institute, Cary, NC). Because the study used data that could not be linked to identifiable women or infants, the Office of Human Subjects Research of the National Institutes of Health granted it an exemption from the requirement for review and approval by the institutional review board.
Maternal and infant characteristics for the analysis group pregnancies and for those of the entire U.S. cohort of mothers aged 15–44 for 1990–2000 are shown in Table 1. For both the young, nulliparous women and the entire U.S. cohort, twins were born at gestational ages 4 to 5 weeks earlier and birth weights 1,000 to 1,100 g smaller than singletons during the study interval. Among the 16- to 19-year-old study group, there were a total of 25,065 twin pregnancies and 3,362,245 singleton pregnancies, giving a total crude twin pregnancy rate of 7.4 per 1,000 pregnancies during the 11-year study period. The total crude twin pregnancy rate in the United States during the same period was 13.1 per 1,000.
Twin gestation rates increased between 1990 and 2000 (Fig. 1). The crude twin pregnancy rate in the group of interest (16- to 19-year-old nulliparous women) increased 13.9%, from 7.2 to 8.2 per 1,000 pregnancies from 1990 to 2000. Among older nulliparous women (aged 20 to 44), there was a much greater increase—43.8%—from 12.8 to 18.4 per 1,000 pregnancies, because of, we assume, the increasing use of fertility treatments by older women. By comparison, the entire population of U.S. women (all ages 15–44, all parity, all races) had a 38.4% increase, from 11.2 to 15.5 per 1,000 pregnancies.
We compared the pattern of changes over the course of time in twin gestation rates before fortification to the pattern after fortification. In the analysis group of 16- to 19-year-old nulliparous women, before fortification, there was a nonsignificant increase in twin gestation rates over the course of time (slope = 0.0052, standard error 0.0034, P = .13), ie, twin rates were essentially constant between January 1990 and November 1996 (Fig. 2).
After fortification, from December 1996 through December 2000, there was a small but statistically significant continuously increasing trend in twinning rates (difference in slope = 0.024, standard error 0.0088, P = .006), indicating that twin gestation rates have increased since the addition of folic acid to the food supply (Fig. 2). Regression analysis showed that since fortification, nulliparous white, black, and Hispanic women aged 16–19, in aggregate, have had a 2.4% (95% CI 0.1–4.2%) per year increase in twin gestation rates. This translates into 2 additional sets of twins per 10,000 deliveries per year.
We examined the influence of folic acid fortification on twin gestation rates by looking for changes in rates among nulliparous 16- to 19-year-old women, a group that is almost never exposed to fertility treatments. If the effect of folic acid fortification were to increase twinning rates, one would expect to see an increase in the slope of the line describing the trend in twin rates after fortification was in effect. We detected a small but statistically significant increase in twin gestation rates among women who conceived after the introduction of folic acid into the food supply. However, the pattern of the increase is inconsistent with a fortification effect.
After several months of exposure to folic acid, serum and red cell folate levels reach a plateau.7,27 If fortification were increasing twin gestation rates, one would expect to have observed a corresponding plateau in twin rates when folate levels reached their steady state in the childbearing population. Such a plateau in twin rates would certainly have been evident by early 2000, because folate stores in the entire maternal population were maximized by late 1998.7 In contrast, we observed a persistent increase that continued after the maximal fortification effect on folate status was reached.
A number of other unmeasured factors may explain the increase in twin gestation rates observed in the late 1990s among young women who were not exposed to fertility treatments. There may have been changes in the accuracy or regularity of reporting infant plurality, particularly among stillbirths.28 Additionally, the use of technology in vital record collection, transmittal, and quality control increased during our study period and facilitated the correction of records, resolution of missing data, and improved reporting accuracy.29–31 It is conceivable that the small increase in twinning we observed is a reflection of better vital records reporting. Finally, a folate-related decrease in NTDs may contribute to improved survival of a small number of twin pairs.
Previous studies of the impact of folic acid on twin gestations have had mixed results. Two international studies32,33 found no increase in twinning with the use of folic acid at doses of 400 μg to 4 mg per day. Two additional studies have examined the impact of the U.S. fortification program. In a population-based study in Texas, Waller and colleagues16 reported continuation of a 1–4% per year increase in twinning that began before fortification. In a similar study in California, Shaw and colleagues17 found no significant increase in twinning prevalence associated with fortification. Each of these studies, however, included in their “prefortification” groups women whose pregnancies had been exposed to folic acid fortification during the optional fortification period beginning March 1996. It is likely that a substantial amount of folic acid entered the food supply after fortification was authorized in 1996 but before the January 1998 deadline. For example, Jacques et al34 reported that most targeted products in New England were fortified by July 1997. The inclusion of women who were exposed to some folic acid in the unexposed population could have produced an underestimate of the effect of folic acid on twinning.
Czeizel et al11 and Ericson et al12 reported significant increases in twinning in women taking folic acid. However, they counted twin births, which overestimates the rate of twin gestation. When Czeizel et al's11 data are reevaluated counting the number of twin pregnancies, the increased twinning rate is no longer statistically significant (44 of 2,198 compared with 29 of 2,170; relative risk 1.50, 95% CI 0.94–2.38; P = .09). Moreover, Czeizel et al's 1994 results are not statistically significant when women who had ovulation induction were excluded.11 When Ericson et al12 eliminated women with a history of involuntary childlessness, folic acid use was still associated with twin births (odds ratio 1.45, 95% CI 1.06–1.98), although it is not clear that excluding women with a history of involuntary childlessness eliminated all women receiving fertility treatments.
Many limitations to our study exist. Because reported data on ARTs do not include specific maternal age, race/ethnicity, or parity, it is not possible to exclude specific ART conceptions from the pooled national data. Patterns of use of ovulation induction medications are not recorded systematically. Therefore, we cannot specifically account for the contribution of fertility treatments to national twin conception rates. We studied 16- to 19-year-old women to avoid confounding by fertility treatments. Between 1997 and 2000, 0.0% of infants conceived by ARTs were born to women younger than 20.22 We cannot exclude the possibility that a twinning response to fortification may be different among women aged 20 years or more, although we know of no data suggesting such a difference.
It is possible that women are exposed to additional folic acid in the form of supplements. Since the 1992 United States Public Health Service recommendation that women capable of becoming pregnant consume 400 μg of folic acid daily, numerous public health education campaigns have attempted to increase the proportion of women using daily supplements. Although awareness of folic acid has increased significantly, only a minority (30% in 1997, 31% in 2002) of women of childbearing age takes supplements regularly.4 Furthermore, teenagers are least likely to be aware of folic acid35 and are least likely to consume daily supplements (22% of ages 18–24, 36–37% of ages 25–45).4
Our study has several major strengths. It is nationally representative, population-based, and includes all 3.4 million twin and singleton births occurring in 16- to 19-year-old women in the United States from 1990 to 2000. Because of the large number of subjects, we were able to take into account the effects of age and race and limit the analysis to nulliparous women and still have sufficient data to detect even small shifts in twin gestation rates. Use of national data eliminates the potential problem that some parts of the country may have been exposed to fortified food before others. A search of PubMed (National Library of Medicine) from 1966 to September 2004 for the terms “twins,” “twinning,” or “multiple pregnancy” and “folic acid” or “fortified food” showed no other studies analyzing twin rates based on the actual date that fortified food was first introduced in the United States during March 1996.
Our data show that in the entire United States, twin gestation rates in young women who were unexposed to fertility treatments have been slowly increasing since the introduction of folic acid fortification of grain. The increase is small, ie, 2 twin pairs per 10,000 deliveries per year in our study population, which is far less the 40% increase reported in those studies reporting that folic acid-containing multivitamins increased twin rates. Moreover, the increase is unlikely to be the result of food fortification because folic acid causes an increase in body folate stores that plateaus within months, and the twin rate continued to increase. Thus, fortification with folic acid does not appear to be causing a serious public health problem related to twinning.
1. Czeizel AE, Dudas I. Prevention of the first occurrence of neural tube defects by periconceptional vitamin supplementation. N Engl J Med 1992;327:1832–5.
2. Prevention of neural tube defects: results of the Medical Research Council Vitamin Study. MRC Vitamin Study Research Group. Lancet 1991;338:131–7.
3. Centers for Disease Control. Recommendations for the use of folic acid to reduce the number of cases of spina bifida and other neural tube defects [review]. MMWR Recomm Rep 1992;41(RR-14):1–7.
4. March of Dimes. Folic acid and the prevention of birth defects: a national survey of pre-pregnancy awareness and behavior among women of childbearing age, 1995–2002. Washington, DC: March of Dimes Foundation; 2002.
5. Food and Drug Administration. Food standards: amendment of standards of identity for enriched grain products to require addition of folic acid. Federal Register 1996;61:8781–97.
6. Centers for Disease Control. Folate status in women of childbearing age, by race/ethnicity–United States, 1999–2000. MMWR Morb Mortal Wkly Rep 2002;51:808–10.
7. Ray JG. Folic acid food fortification in Canada. Nutr Rev 2004;62(suppl):S35–9.
8. Honein MA, Paulozzi LJ, Mathews TJ, Erickson JD, Wong LY. Impact of folic acid fortification of the U.S. food supply on the occurrence of neural tube defects [published erratum appears in JAMA 2001;286:2236]. JAMA 2001;285:3022–3.
9. Centers for Disease Control (CDC). Spina bifida and anencephaly before and after folic acid mandate: United States, 1995–1996 and 1999–2000. MMWR Morb Mortal Wkly Rep 2004;53:362–5.
10. Mills JL, Signore C. Neural tube defect rates before and after food fortification with folic acid. Birth Defects Res A Clin Mol Teratol 2004;70:844–5.
11. Czeizel AE, Metneki J, Dudas I. The higher rate of multiple births after periconceptional multivitamin supplementation: an analysis of causes. Acta Genet Med Gemellol (Roma) 1994;43:175–84.
12. Ericson A, Källén B, Åberg A. Use of multivitamins and folic acid in early pregnancy and multiple births in Sweden. Twin Res 2001;4:63–6.
13. Werler MM, Cragan JD, Wasserman CR, Shaw GM, Erickson JD, Mitchell AA. Multivitamin supplementation and multiple births. Am J Med Genet 1997;71:93–6.
14. Martin JA, Park MM. Trends in twin and triplet births: 1980–97. Natl Vital Stat Reps 1999;47:1–16.
15. From the Centers for Disease Control and Prevention. Use of assisted reproductive technology–United States, 1996 and 1998. MMWR Morb Mortal Wkly Rep 2002;51:97–101.
16. Waller DK, Tita AT, Annegers JF. Rates of twinning before and after fortification of foods in the U.S. with folic acid, Texas, 1996 to 1998. Paediatr Perinat Epidemiol 2003;17:378–83.
17. Shaw GM, Carmichael SL, Nelson V, Selvin S, Schaffer DM. Food fortification with folic acid and twinning among California infants. Am J Med Genet A 2003;119:137–40.
18. National Center for Health Statistics. Linked birth/infant death public use CD-ROMs; 1990 and 1991 birth cohorts. Hyattsville (MD): Public Health Service; 1996.
19. National Center for Health Statistics. Natality public use CD-ROMs; 1992–2000 natality files. Hyattsville (MD): Public Health Service; 1996–2002.
20. National Center for Health Statistics. Fetal death public use tapes and CD-ROM; 1990–1994. Hyattsville (MD): Public Health Service; 1994–1996.
21. National Center for Health Statistics. Perinatal mortality public use CD-ROMs; 1995–2000 numerator files. Hyattsville (MD): Public Health Service; 1998–2003.
22. Reynolds MA, Schieve LA, Martin JA, Jeng G, Macaluso M. Trends in multiple births conceived using assisted reproductive technology, United States, 1997–2000. Pediatrics 2003;111:1159–62.
23. Mosher WD, Pratt WF. Fecundity and infertility in the United States, 1965–88: advance data from vital and health statistics; no. 192. Hyattsville (MD): National Center for Health Statistics; 1990.
24. Wilcox LS, Mosher WD. Use of infertility services in the United States. Obstet Gynecol 1993;82:122–7.
25. Contribution of assisted reproductive technology and ovulation-inducing drugs to triplet and higher-order multiple births–United States, 1980–1997. MMWR Morb Mortal Wkly Rep 2000;49:535–8.
26. McCullagh P, Nelder JA. Generalized linear models. 2nd ed. London (UK): Chapman and Hall; 1989.
27. Caudill MA, Cruz AC, Gregory JF 3rd, Hutson AD, Bailey LB. Folate status response to controlled folate intake in pregnant women. J Nutr 1997;127:2363–70.
28. Centers for Disease Control (CDC). Racial/ethnic trends in fetal mortality–United States, 1990–2000. MMWR Morb Mortal Wkly Rep 2004;53:529–32.
29. Starr P, Starr S. Reinventing vital statistics: the impact of changes in information technology, welfare policy, and health care. Public Health Rep 1995;110:534–44.
31. National Center for Health Statistics. Technical appendix: 2000 natality detail file. NCHS CD-ROM Series 21, No. 13H. Hyattsville (MD): U.S. Department of Health and Human Services, Centers for Disease Control and Prevention; 2002.
32. Mathews F, Murphy M, Wald NJ, Hackshaw A. Twinning and folic acid use. Lancet 1999;353:291–2.
33. Li Z, Gindler J, Wang H, Berry RJ, Li S, Correa A, et al. Folic acid supplements during early pregnancy and likelihood of multiple births: a population-based cohort study. Lancet 2003;361:380–4.
34. Jacques PF, Selhub J, Bostom AG, Wilson PW, Rosenberg IH. The effect of folic acid fortification on plasma folate and total homocysteine concentrations. N Engl J Med 1999;340:1449–54.
35. Alozie Arole CN, Puder KS, Reznar M, Eby E, Zhu BP. Folic acid awareness in Michigan, 1996–1999. Obstet Gynecol 2003;102:1046–50.