| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the *Division of Maternal–Fetal Medicine and Section of Epidemiology and Biostatistics, Department of Obstetrics, Gynecology and Reproductive Sciences, University of Medicine and Dentistry of New Jersey–Robert Wood Johnson Medical School/Robert Wood Johnson University Hospital, New Brunswick, New Jersey.
Address reprint requests to: John C. Smulian, MD, MPH, UMDNJ–Robert Wood Johnson Medical School, 125 Paterson Street, CAB-2140, New Brunswick, NJ 08901-1977; e-mail: smuliajc{at}umdnj.edu.
 |
ABSTRACT |
|---|
TOP
 ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS: United States natality data were used to assess trends in twin pregnancies resulting in live births. This age-period-cohort analysis included 7, 5-year maternal-age groups (15–19 through 45–49 years), 6 twin delivery periods (1975, 1980, 1985, 1990, 1995, and 2000), and 12, 5-year maternal birth cohorts (1926–1930 through 1981–1985). The independent effects of maternal age, twin delivery period, and maternal birth cohort on twin delivery rates for blacks and whites were modeled using Poisson regression techniques.
RESULTS: Our study assessed 95,042 blacks and 401,989 whites with twin deliveries. Twin deliveries increased by 46% for blacks and 62% for whites from 1975 to 2000, with the largest increase occurring in the year 2000. For blacks, maternal age had the strongest impact on the increasing twin delivery rates, followed by period of delivery. For whites, the greatest effect was due to period of delivery, followed by maternal birth year cohort and, lastly, maternal age.
CONCLUSION: Our data confirm the importance of nature's biologic contribution of maternal aging to twin delivery rates, but suggest that recent changes in the environment surrounding pregnancy (nurture) also influence twin delivery rates. The relative contributions of biologic versus environmental influences appear to differ among blacks and whites.
LEVEL OF EVIDENCE: II-2
A statistical methodology known as age-period-cohort analysis can be used to examine 3 closely related time factors (maternal age, period of delivery, and maternal birth cohort) to determine independent effects of these factors on outcome.10 This analytic approach can produce some insight from an epidemiologic perspective on interactions between nature and nurture, where maternal age represents biologic effects of nature and period (year) of birth and maternal birth-year cohort represents environmental influences, or nurture. Two of these, age and period of birth, are frequently the object of study, but are seldom examined together. It is not known whether the effect of maternal age on twin deliveries is consistent across time periods. Significantly, period-of-birth studies that examine trends in twin birth rates usually do not consider how those trends differ by maternal birth cohort. Examination of maternal birth-cohort effects might provide clues as to whether birth or early childhood events influence reproductive health. In addition, because age and gravidity are closely linked, the aging effect of the uterus (repeated pregnancies) has implications for any analysis related to trends over time. Maternal age is particularly important when examining issues related to birth cohorts to separate the effects of biologic age from potential environmental influences experienced early in life.
Because time is such an important variable in understanding twin birth trends, we designed this study to assess the influences of maternal age, period (year) of delivery, and maternal birth cohort on trends in the rates of twin deliveries in the United States. We hypothesized that each of these factors would exert an effect on twin birth rates, but that period of delivery would have the strongest effect concurrent with iatrogenic causes of multiple gestations. We also examined whether these effects differed between blacks and whites.
|
MATERIALS AND METHODS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
To study trends in twin pregnancy deliveries rather than twin liveborn infants, the total number of twin pregnancies delivering was estimated by dividing the number of twin liveborn infants in half. This prevented mothers from being counted twice (once for each infant born). Twin pregnancy delivery rates, therefore, were estimated by this formula:
|
| (1) |
We examined 7, 5-year maternal age groups (15–19 years through 45–49 years), 6 twin delivery periods (1975, 1980, 1985, 1990, 1995, 2000), and 12, 5-year maternal birth cohorts (1926–1930 through 1981–1985). We excluded births to women less than 15 or more than 49 years of age because the number of twin births in these extreme age groups in the earlier periods were too few.
The independent effects of maternal age, delivery period, and maternal birth cohort on twin delivery rates were modeled using Poisson regression techniques. Because any 2 age-period-cohort factors are sufficient to derive the third (cohort = delivery period – age), this introduces an exact linear dependency among these 3 effects in the regression models. To overcome this (nonidentifiability) limitation, we constrained the effect of the last maternal birth cohort (ie, 1981–85) to zero. This allowed maternal age, period, and birth cohorts to be modeled as categorical predictors: maternal age with 7 indicators, period of delivery with 6, and birth-year cohort with 12 indicators in the Poisson regression models. Interactions among the variables were assessed, and model-based variance estimates were corrected for "extra Poisson" variability12 based on the deviance measure. The relative contributions of each of the age-period-cohort factors on twin pregnancy delivery rates was based on the deviance 
2 statistic. The goodness-of-fit of these regression models was assessed based on the overall deviance (2) in relation to the total degrees of freedom. Thus, P 
.05 would indicate reasonable fit, whereas P < .05 would indicate poor fit or lack of fit.
Because maternal race has been implicated as a contributor to twin delivery rates,8,13–16 analyses were performed separately for blacks and whites. Other race and ethnicity subgroups were not examined separately because of insufficient numbers. Disparity between blacks and whites in twin pregnancy delivery rates for all combinations of maternal age, period of delivery, and maternal birth-year cohort were estimated by risk ratios derived from Poisson regression models.
Statistical analysis was performed with SAS 8.2 (SAS Institute Inc, Cary, NC) operating on a UNIX operating system. The study received ethics approval from the Institutional Review Board of the University of Medicine and Dentistry of New Jersey–Robert Wood Johnson Medical School.
|
RESULTS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
When the contributions of maternal age, delivery period, and maternal birth-year cohort to trends in twin deliveries were examined, the patterns were markedly different for blacks and whites (Table 4). For blacks, maternal age had the strongest influence on increasing twin pregnancy delivery rates, followed by the period (year of delivery). Maternal birth cohort contributed little to twin pregnancy deliveries for blacks. For whites, however, the greatest effect was due to period, followed by birth cohort, and, lastly, maternal age. The deviance 2 results are provided separately for blacks and whites and represent the relative importance of age, period, and cohort to twin pregnancy delivery rates within each group. The deviance is not a measure of the absolute magnitude of the effect from each variable. Although age, period, and maternal birth cohort provided satisfactory fit to twin pregnancy delivery rates among blacks (P = .105), the fit of the model to whites was poor (P < .001).
|
Advancing maternal age was associated with gradual increases in twin pregnancy deliveries across all periods for both blacks (Fig. 2A) and whites (Fig. 2B). However, a sharp increase in the rate was evident among whites, aged 40 years and over, consistently across all periods. Because the year 2000 birth period was associated with unexpectedly large increases in twin delivery rates across all maternal ages for both blacks and whites, we performed a subanalysis of the relative contributions of age, period, and cohorts in twin deliveries after excluding the year 2000 period. In this analysis, the period effect for blacks disappeared, leaving only maternal age as a significant contributor to twin pregnancy delivery rates in this group. For whites, the contributions of age, period, and cohort were persistent, with an attenuated period effect.
|
Trends in twin deliveries by maternal age within the strata of maternal birth cohorts are shown in Figure 3. In general, among blacks (Fig. 3A) and whites (Fig. 3B), twin pregnancy delivery rates increased by maternal age within each birth cohort, with the increase being somewhat steeper among whites than among blacks. However, whites born in the earlier cohorts (ie, 1931–55), aged 40 years and over, had the sharpest increase in twin pregnancy delivery rates. The rate was generally higher for blacks than whites for any combination of maternal age, period, and birth cohort except when maternal ages were 45 years or greater.
|
|
DISCUSSION |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Our results also agree with prior observations that the overall rate for twin deliveries is higher among blacks than among whites.13–16 It is not clear why there were much smaller age-related increases among blacks. If twinning rates were primarily influenced by advancing age, similar patterns of twin deliveries should be observed in both blacks and whites. However, the age-related effect for blacks was driven predominantly by the oldest age group (45–49 years), which has the lowest natural fertility rates. Similar to whites, we believe that the spike in rates in this older group is, at least in part, due to the use of various ART strategies.
Disparities between blacks and whites in access to medical services, as well as use of available medical services, have been well described.19,20 Therefore, we speculate that the limited period effect and the age effect (mostly 45–49 years) on twin pregnancy delivery rates for blacks may reflect issues related to the use of medical services. During the time of this study, particularly if delivery was before 2000 and if the women were younger than age 45, blacks may have had either reduced access to ART services or less use of available services. It has been suggested that blacks might have higher clinical pregnancy rates with ART, which may predispose to greater numbers of multiple gestations.21 If this is true and if black women in recent years have begun to increase their use of ART, greater rates of twins in this group may be expected in the future.
The Society for Assisted Reproductive Technology, in collaboration with the United States Centers for Disease Control and Prevention, publish annual clinic-specific success rates for ART procedures in the United States, but the society's data does not routinely include information about either race or other forms of assisted conception. Several investigators have reported that older, white, married women of higher socioeconomic status were most likely to have obtained services related to infertility.22,23 Green et al24 reported that, for women seeking infertility evaluations, approximately 85% were white and 10% black, which was similar to the population of the city where the study was performed. However, their study included subjects from an urban Medicaid infertility center and from a suburban private insurance center, which may reflect biased ascertainment.24 More importantly, more than 25% of blacks had infertility that was related to previous sterilization, compared with less than 5% for whites. In past years those cases may have been more likely to be treated with surgery rather than with ART procedures, which increase the rate of multiple gestations. In a cohort of 13,151 women from a single insurer, Lynch et al25 reported that race did not contribute to multiple births, but only 7.7% were black women, and there were only 244 multiple-gestation births in the entire cohort.
One unexpected finding from this study is the marked period effect for the year 2000 births, which includes all age groups for both blacks and whites. This effect is seen even at younger ages, where use of ART would be uncommon. We believe that this recent rate increase in twin deliveries may be due, in part, to a change in the maternal environment from increased periconceptional folic acid intake. Public health policy changes leading to increased dietary fortification with folate in the mid-1990s and campaigns promoting folic acid supplementation before and during pregnancy were designed to reduce the incidence of neural tube defects. A number of studies have suggested that increased periconceptional folic acid also increases the rate of twin births by as much as 40%.9,26,27 Using modeling techniques, Lumley et al9 estimated that, for every 100,000 pregnancies at or beyond 20 weeks of gestation, there would be an additional 572 twin confinements because of periconceptional folate, with their associated adverse perinatal outcome risks. Interestingly, this effect appears mainly for dizygotic twinning26,28 and may be related to improved early pregnancy survival of multiple gestations with increased folate levels.9 This is supported by the observation that mothers with a methylenetetrahydrofolate reductase gene mutation have been reported to have a 2.28-fold lower rate of twin pregnancies.27 Of note, this mutation is associated with mildly elevated homocyst(e)ine levels in the presence of lower folate levels.
The major strengths of this analysis are the large numbers of subjects available for inclusion and the population-based nature of the data. Nevertheless, our method for calculation of twin pregnancy delivery rates may have underestimated the actual numbers of twin pregnancies because those pregnancies with a demise of one or both fetuses would not be fully captured. The impact of this is likely minimal, given the large number of subjects. Although these data do not include information about periconceptional folic acid intake or use of ART to achieve pregnancy, it is reasonable to believe that these are important factors in explaining our observed trends.
Our data suggest that intrinsic biologic (nature) factors (maternal age and race) exert a dominant effect on higher twin pregnancy delivery rates among blacks and that environmental (nurture) influences (birth period and maternal birth-year cohort) are the most important drivers for twin pregnancy delivery rates among whites. These results raise intriguing questions about past differences in access to and use of reproductive medical services that might contribute to disparities in twin pregnancy delivery rates. These results also demonstrate an apparently recent exposure among both blacks and whites that has increased rates. Because folic acid supplementation is a candidate contributor to this observation, we recommend future studies be designed to specifically examine this potential interaction. Identifying the contributors to increasing twin delivery rates should ultimately lead to balanced strategies to moderate these rates.
|
Footnotes |
|---|
Received January 8, 2004. Received in revised form April 2, 2004. Accepted May 17, 2004.
10.1097/01.AOG.0000134524.58795.bd
|
REFERENCES |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2. Kinzler WL, Ananth CV, Vintzileos AM. Medical and economic effects of twin gestations. J Soc Gynecol Investig 2000;7:321–7.[Medline]
3. Keith LG, Oleszczuk JJ, Keith DM. Multiple gestation: reflections on epidemiology, causes, and consequences. Int J Fertil Womens Med 2000;45:206–14.[Medline]
4. Zhang J, Meikle S, Grainger DA, Trumble A. Multifetal pregnancy in older women and perinatal outcomes. Fertil Steril 2002;78:562–8.[Medline]
5. 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.
6. Jewell SE, Yip R. Increasing trends in plural births in the United States. Obstet Gynecol 1995;85:229–32.[Abstract]
7. Kiely JL, Kiely M. Epidemiological trends in multiple births in the United States, 1971–1998. Twin Res 2001;4:131–3.[Medline]
8. Russell RB, Petrini JR, Damus K, Mattison DR, Schwarz RH. The changing epidemiology of multiple births in the United States. Obstet Gynecol 2003;101:129–35.
9. Lumley J, Watson L, Watson M, Bower C. Modelling the potential impact of population-wide periconceptional folate/multivitamin supplementation on multiple births. BJOG 2001;108:937–42.[Medline]
10. Holford TR. Understanding the effects of age, period, and cohort on incidence and mortality rates. Annu Rev Public Health 1991;12:425–57.[Medline]
11. MacDorman MF, Atkinson JO. Infant mortality statistics from the linked birth/infant death data set-1995 period data. Monthly vital statistics report; Vol 46, no. 6, suppl 2. Hyattsville (MD): National Center for Health Statistics, 1998.
12. Liang KY, McCullaugh P. Case studies on binary dispersion. Biometrics 1993;49:623–30.[Medline]
13. Mushinski M. Trends in multiple births. Stat Bull Metrop Insur Co 1994;75:28–35.[Medline]
14. Pollard R. Ethnic comparison of twinning rates in California. Hum Biol 1995;67:921–31.[Medline]
15. Martin JA, Taffel SM. Current and future impact of rising multiple birth ratios on low birthweight. Stat Bull Metrop Insur Co 1995;76:10–8.[Medline]
16. Oleszczuk JJ, Cervantes A, Kiely JL, Keith DM, Keith LG. Maternal race/ethnicity and twinning rates in the United States, 1989–1991. J Reprod Med 2001;46:550–7.[Medline]
17. Callahan TL, Hall JE, Ettner SL, Christiansen CL, Greene MF, Crowley WF. The economic impact of multiple-gestation pregnancies and the contribution of assisted-reproduction techniques to their incidence. N Engl J Med 1994;331:244–9.
18. Frank O, Bianchi PG, Campana A. The end of fertility: age, fecundity and fecundability in women. J Biosoc Sci 1994;26:349–68.[Medline]
19. Fiscella K, Franks P, Doescher MP, Saver BG. Disparities in health care by race, ethnicity, and language among the insured: findings from a national sample. Med Care 2002;40:52–9.[Medline]
20. Schneider EC, Cleary PD, Zalavsky AM, Epstein AM. Racial disparity in influenza vaccination: does managed care narrow the gap between African-Americans and whites? JAMA 2001;286:1455–60.
21. Nichols JE Jr, Higdon HL 3rd, Crane MM 4th, Bone WR. Comparison of implantation and pregnancy rates in African-American and white women in an assisted reproductive technology practice. Fertil Steril 2001;76:80–4.[Medline]
22. Wilcox LS, Mosher WD. Use of infertility services in the United States. Obstet Gynecol 1993;82:122–7.
23. Hirsch MB, Mosher WD. Characteristics of infertile women in the United States and their use of infertility services. Fertil Steril 1987;47:618–25.[Medline]
24. Green J, Robins JC, Scheiber M, Awadalla S, Thomas M. Racial and economic demographics of couples seeking infertility treatment. Am J Obstet Gynecol 2001;184:1080–2.[Medline]
25. Lynch A, McDuffie R Jr, Murphy J, Faber K, Leff M, Orleans M. Assisted reproductive interventions and multiple birth. Obstet Gynecol 2001;97:195–200.
26. Ericson A, Kallen B, Aberg A. Use of multivitamins and folic acid in early pregnancy and multiple births in Sweden. Twin Res 2001;4:63–6.[Medline]
27. Czeizel AE, Metneki J, Dudas I. The higher rate of multiple births after periconceptional multivitamin supplementation: an analysis of causes. Acta Genet Med Gemellol 1994;43:175–84.[Medline]
28. Hasbargen U, Lohse P, Thaler CJ. The number of dichorionic twin pregnancies is reduced by the common MTHFR 677C
T mutation. Hum Reprod 2000;15:2659–62.
This article has been cited by other articles:
|
|
 |
|
  C. A. Mathews, B. Bimson, T. L. Lowe, L. D. Herrera, C. L. Budman, G. Erenberg, A. Naarden, R. D. Bruun, N. B. Freimer, and V. I. Reus Association Between Maternal Smoking and Increased Symptom Severity in Tourette's Syndrome Am J Psychiatry, June 1, 2006; 163(6): 1066 - 1073. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
