An increasing number of women in developing countries are postponing pregnancy.1 Previous studies have shown associations between advanced maternal age and a higher risk of miscarriage and chromosomal abnormalities of the fetus,2–8 whereas studies on the associations with congenital malformations, preterm birth, and stillbirth have reported conflicting findings.3,9–17
Many previous studies have been inadequately powered to explore associations with rare adverse outcomes or to adjust for potential important confounders. Furthermore, most studies examining rare fetal outcomes base their results on the prevalence of adverse outcomes at 20–40 weeks of gestation or on live birth prevalence.4–7,8,16 Thus, such studies do not include fetal anomalies, which result in fetal loss before 20 weeks of gestation. Moreover, some previous studies have investigated associations with maternal ages and adverse pregnancy outcomes in a selected group of pregnant women undergoing invasive testing or those willing to pay for antenatal care services, making their results more prone to bias.
The objective of this study was to examine the possible associations between advanced maternal age and adverse pregnancy outcomes in a large population-based cohort of pregnant women in Denmark followed from the first trimester to delivery or termination of pregnancy. This would provide more comprehensive and accurate risk estimates of fetal anomalies detected in early pregnancy. An additional aim was to predict risk estimates of a composite adverse pregnancy outcome to identify whether certain groups of pregnant women were at particular risk of the selected adverse pregnancy outcomes.
MATERIALS AND METHODS
Our source population consisted of pregnancies with a first-trimester screening performed between 11 and 14 weeks of gestation at one of the public hospitals in Denmark between January 2008 and December 2014.
Since 2004, Denmark has had a national pregnancy screening program that offers first-trimester screening for chromosomal abnormalities at 11–14 weeks of gestation and a second-trimester screening for fetal anomalies at 18–20 weeks of gestation to all pregnant women.18 The screening and follow-up programs for pregnant women in Denmark do not differ according to maternal age. More than 95% of Danish women choose to have a first-trimester screening performed.19 In this study, we included singleton pregnancies in women aged 20 years and older.
We retrieved data from the Danish Fetal Medicine Database, which was established in 2008.19 Information on maternal characteristics such as age at time of participation in the first-trimester ultrasound scan, smoking status, body mass index (BMI, calculated as weight (kg)/[height (m)]2), ethnicity, parity, use of assisted reproductive technology (ART), and information from the ultrasound screenings are continuously received from local servers at all departments of gynecology and obstetrics in Denmark.19 Furthermore, the database collects information about outcome of the pregnancy, including fetal losses and postnatally detected congenital malformations from the National Patient Registry, information about pregnancy complications and delivery from the Medical Birth Registry, and results from prenatal and postnatal chromosome analyses from the Danish Cytogenetic Registry.19 The Danish Fetal Medicine Database holds information on pregnancy outcomes for more than 95% of all pregnancies in Denmark.19
We defined advanced maternal age as maternal age 35 years or older at the time of first-trimester screening. These women were further categorized into two subgroups: 35–39 years and 40 years or older. The reference group was defined as pregnant women between 20 and 34 years of age.
An adverse pregnancy outcome was defined as the occurrence of chromosomal abnormalities, congenital malformations, miscarriage, stillbirth, or birth before 34 weeks of gestation (see further definitions in Table 1). Chromosomal abnormalities and congenital malformations include both prenatally and postnatally detected cases.
For congenital malformations, we excluded minor congenital malformations according to the European Surveillance of congenital anomalies exclusion guidelines.20 In addition, we reviewed the International Classification of Diseases, 10th Revision (ICD-10) list of congenital malformations and excluded cases with irrelevant or less severe recordings such as “increased nuchal translucency” and “patent ductus arteriosus.” We included cases with unspecific ICD-10 codes, because the severity of the malformation could not be further assessed. Such cases were “hydronephrosis” (n=1,388), which is an ICD-10 code often given prenatally in cases with also mild dilatation (less than 10 mm) of the renal pelvis.
All five adverse pregnancy outcomes were formed into a single composite outcome. In the composite outcome, an adverse pregnancy outcome only counted as one event, although the same woman might have had more than one of the adverse pregnancy outcomes during the same pregnancy.
The potential confounders of possible associations between advanced maternal age and risk of adverse pregnancy outcomes were defined a priori based on the literature.21–27 The potential confounders included in this study were smoking, maternal ethnicity, maternal BMI, parity, use of ART, and pregnancy history with a previous chromosomal abnormality. Smoking and maternal BMI were self-reported at the time of first-trimester screening.
Analyses were performed using STATA 14. Differences between maternal age groups according to demographic variables and prevalence of adverse pregnancy outcomes were investigated using Pearson χ2 tests.
To explore the potential associations between advanced maternal age and risk of adverse pregnancy outcomes, multivariable logistic regression analyses were performed comparing each of the two advanced maternal age groups with the reference. This was carried out for each of the five adverse pregnancy outcomes and the composite outcome. We dealt with the potential problem of multiple testing using the Bonferroni correction with an α level of 0.002. Therefore, the results of the multivariable logistic regression analyses were presented as both crude and adjusted odds ratio (OR) estimates with 99.8% CIs.
When analyzing the association between advanced maternal age and risk of miscarriage, stillbirth, and birth before 34 weeks of gestation, we excluded pregnancies with prenatally detected chromosomal abnormalities and congenital malformations. Sensitivity analyses were performed to assess the extent to which the association between advanced maternal age and risk of congenital malformations were affected by changes in the inclusion or exclusion of minor or unspecific malformation types.
Because some prior studies have included use of ART in the multiple regression analyses as a potential confounding factors, whereas other prior studies have excluded this factor as a result of an assumed intermediate factor, we did a subanalysis of the multiple logistic regression model where use of ART was excluded to check whether the associations changed significantly. Another subanalysis was performed to test whether the associations changed significantly when including pregnant women with a missing parity registration in the logistic regression analyses.
To estimate whether groups of pregnant women with specific combinations of certain risk predictors had a higher risk of experiencing an adverse pregnancy outcome, a risk prediction chart was made with the composite outcome. The following predictors were chosen a priori and included in the model: maternal age, type of conception, parity, BMI, and smoking status.21,24–27 We assumed that there were no interactions between the chosen risk predicting factors corresponding to the current literature. The risk prediction chart was constructed by predicting the absolute risk for each combination of the five predictors in the fitted logistic regression model.28 In total, 144 different combinations of obstetric and maternal factors were estimated to predict risk scores of a composite pregnancy outcome. The log odds were converted to risk estimates using the formula: .28
The study was approved by the Danish Data Protection Agency (j.nr. 2017-41-5057) and the Danish Fetal Medicine Database (ref. no. FØTO-2017-02-16).
The study population consisted of 369,516 singleton pregnancies in women aged 20 years or older with a first-trimester screening performed. The average age of the pregnant women in our study population was 30.1 years.
Table 2 shows maternal and obstetric characteristics. The three maternal age groups differed with regards to BMI, ethnicity, parity, use of ART, and a pregnancy history with chromosomal abnormalities. Pregnant women in both advanced maternal age groups were more likely to have slightly higher BMIs, to have conceived after ART, and to have pregnancy histories with chromosomal abnormalities, whereas they were less likely to be smokers, of Caucasian ethnicity, and less often nulliparous when compared with the reference group.
The prevalence of adverse pregnancy outcomes according to maternal age is shown in Table 3. For all age groups, congenital malformations were the most frequent adverse pregnancy outcome, whereas stillbirth was the rarest observed adverse event in the study population. The prevalence of adverse pregnancy outcomes differed among the maternal age groups, in which all of the selected adverse pregnancy outcomes occurred more often in the two advanced maternal age groups than in the reference group of pregnant women aged 20–34 years.
Table 4 shows findings from the multivariable logistic regression analyses while accounting for multiple comparison by applying an α level of 0.002. Compared with crude regression analyses, the associations did not change markedly after adjusting for BMI, ethnicity, smoking, use of ART, and a pregnancy history with chromosomal abnormalities in the multivariable regression analyses.
The adjusted regression analyses show that pregnant women in the two advanced maternal age groups had a higher risk of chromosomal abnormalities, miscarriage, and birth before 34 weeks of gestation compared with younger maternal ages. When pregnant women aged 40 years or older were compared with pregnant women aged 20–34 years, they were at a higher risk of having a fetus with chromosomal abnormalities (3.83% vs 0.56%, OR 7.44 [99.8% CI 5.93–9.34]), miscarriage (1.68% vs 0.42%, OR 3.10 [99.8% CI 2.19–4.38]), and giving birth before 34 weeks of gestation (2.01% vs 1.21%, OR 1.66 [99.8% CI 1.23–2.24]). The risk of stillbirth was higher for maternal ages 35–39 years (0.35% vs 0.28%, OR 1.43 [99.8% C: 1.05–1.96]), but no statistically significant difference was observed for pregnant women aged 40 years or older (0.43% vs 0.28%, OR 1.47 [99.8% CI 0.76–2.84]) relative to 20- to 34-year-old pregnant women. There was no difference in risk of congenital malformations among the women 35–39 years of age (3.49% vs 3.32%, OR 1.02 [99.8% CI 0.92–1.12]) and the women 40 years of age or older (3.93% vs 3.32%, OR 1.18 [99.8% CI 0.97–1.44]) who were pregnant relative to the younger maternal age group.
The composite endpoint showed that among the 9,743 pregnant women aged 40 years or older, 1,054 (10.82%) experienced one or more of the selected adverse pregnancy outcomes, and their risk was more than twice (OR 2.02 [99.8% CI 1.78–2.29]) as high when compared with the 300,863 pregnant women aged 20–34 years, of whom 16,429 (5.46%) experienced a selected adverse pregnancy outcome.
A subanalysis in which we excluded ART in the multivariable logistic regression analyses in contrast to including it as a potential confounder did not change the associations notably (data not shown). A subanalysis with the inclusion of pregnant women with a missing parity registration did not change the associations we found significantly (data not shown).
A risk prediction chart is presented in Appendix 1 (available online at http://links.lww.com/AOG/B67) showing estimated risks of experiencing an adverse pregnancy outcome for different age groups of pregnant women according to different levels of the a priori chosen risk predictors: maternal age, parity, mode of conception, BMI, and smoking status. Colors are used to illustrate the level of risk score in the prediction chart. Clarification of calculations and assumptions behind the risk prediction chart appears in Appendix 2 (available online at http://links.lww.com/AOG/B67).
The predictive risk of experiencing an adverse pregnancy outcome was higher for groups of pregnant women at advanced maternal age. Furthermore, the risk prediction chart showed that smoking during pregnancy, nulliparous pregnancy, having conceived after ART, and being overweight or obese predicted a higher risk of an adverse pregnancy outcome across all maternal age groups.
The group of pregnant women aged 40 years or older, who smoked during pregnancy, were obese, were first-time pregnant, and conceived after ART resulted in the highest predictive risk of 20% (CI 18–22%) of experiencing one or more of the selected adverse pregnancy outcomes. In contrast, the group of pregnant women aged 20–34 years, who were also smokers, obese, first-time pregnant, and conceived after ART technology, had a predictive risk of 11% (CI 10–12%), illustrating that maternal age drives a high proportion of the total risk score.
Our findings supported the well-known association between advanced maternal age and substantially higher risk of chromosomal abnormalities and miscarriage4–8 potentially caused by aging processes in the ovaries increasing the rate of meiotic nondisjunction.10 Although previous studies have reported contradictory findings, we found that women at advanced maternal ages were at higher risk of giving birth before 34 weeks of gestation.3 This could potentially reflect obstetric complications known to increase with higher maternal ages. We found a higher risk of stillbirth among pregnant women 35–39 years of age, whereas no difference was found for those 40 years or older. This counterintuitive finding could possibly occur because induction of labor for comorbidities is more likely in women older than 40 years. Risk of congenital malformations did not show any differences among the maternal age groups, as consistent with previous studies.3,14,15 Such adverse pregnancy outcomes could potentially be caused by increased uterine and placental dysfunction, environmental and behavioral exposures, and comorbidities, which are likely to accumulate with advancing maternal age.10 Still, the etiology of most adverse pregnancy outcomes remains unknown and requires further investigation.10
A considerably novel aspect of this study is the comprehensive range of adverse pregnancy outcomes studies in a population-based cohort with a very high national uptake of antenatal screening using registry linkage of rigorously collected data. In this study, we were able to include detected chromosomal abnormalities and congenital malformations from an early gestational age, which in other studies would not have been available for assessment as a result of spontaneous losses or terminations of pregnancies.8,14,16,17 Therefore, this study provides a more accurate risk estimate of adverse pregnancy outcomes already in a first-trimester setting, although miscarriages occurring before the pregnant women participate in the first-trimester screening are not included.
This study was based on prospectively collected information on pregnancy history and outcome with virtually no loss to follow-up (less than 5%). Because maternal age was registered using the unique Civil Registration Number, the exposure did not suffer from misclassification. We defined maternal age at the time of first-trimester screening, which is important to consider when comparing it with other studies that usually defines maternal age at delivery. We assumed no false-positive prenatally detected cases or false-negative postnatally detected cases of chromosomal abnormalities and congenital malformations occurred. Moreover, the severity of the adverse pregnancy outcomes and the registrations using the ICD-10 made it unlikely that outcomes in our study were differentially misclassified among maternal age groups.
Sensitivity analyses that assessed the extent to which the association between advanced maternal age and risk of congenital malformations were affected by changes in the inclusion or exclusion of minor or unspecific malformation types such as “increased nuchal translucency” or “hydronephrosis” did not significantly affect the association we found between advanced maternal age and risk of congenital malformations. We were not able to adjust for all potential confounders assumed to bias the associations between maternal age and adverse pregnancy outcomes such as maternal comorbidities and obstetric factors resulting from inaccessibility or registrations of poorer quality.
The study design allowed us to explore multiple adverse pregnancy outcomes. The composite outcome gives insight into the total risk of selected adverse pregnancy outcomes, although equalizing the importance of the five outcomes might be problematic because of different risks and severity. The high participation rate of the first-trimester screening in Denmark contributed to a highly representative sample of adverse pregnancy outcomes in Danish women. The results can be generalized to other populations with similar antenatal care services, demographic characteristics, and environmental exposures, although the external validity might be challenged by varying prevalence of adverse pregnancy outcomes across countries.
The risk prediction chart offers crucial awareness of which pregnancy groups according to maternal age to consider for closer follow-up during pregnancy. The risk prediction chart cannot function as an accurate patient-specific tool that predicts individual risk as a result of lack of information from other important or more specified risk predictors that are either not fully known such as genetics or factors that we were not able to include in this study such as maternal comorbidities and obstetric factors.
In conclusion, this study found an association between advanced maternal age and higher risk of chromosomal abnormalities, miscarriage, and birth before 34 weeks of gestation. There was an association with stillbirth in women aged 35–39 years, whereas there was no increased risk in women 40 years or older and no increased risk of congenital malformation in the two advanced maternal age groups. Our risk prediction chart may potentially become a tool to identify groups of pregnant women at higher risk of experiencing an adverse pregnancy outcome, leading to increased surveillance, improved clinical counseling, and optimized antenatal care services for these pregnant women.
2. Nybo Andersen A, Wohlfahrt J, Christens P, Olsen J, Melbye M. Maternal age and fetal loss: population based register linkage study. BMJ 2000;320:1708–12.
3. Khalil A, Syngelaki A, Maiz N, Zinevich Y, Nicolaides KH. Maternal age and adverse pregnancy outcome: a cohort study. Ultrasound Obstet Gynecol 2013;42:634–43.
4. Park IY, Kwon JY, Kim YH, Kim M, Shin JC. Maternal age-specific rates of fetal chromosomal abnormalities at 16-20 weeks' gestation in Korean pregnant women >or=35 years of age. Fetal Diagn Ther 2010;27:214–21.
5. Kim JY, Lee JE, Kim SH, Shim SS, Cha DH. Maternal age-specific rates of fetal chromosomal abnormalities in Korean pregnant women of advanced maternal age. Obstet Gynecol Sci 2013;56:160–6.
6. Loane M, Morris JK, Addor MC, Arriola L, Budd J, Doray B, et al. Twenty-year trends in the prevalence of down syndrome and other trisomies in Europe: impact of maternal age and prenatal screening. Eur J Hum Genet 2013;21:27–33.
7. Hartwig T, Sørensen S, Jørgensen FS. The maternal age-related first trimester risks for trisomy 21, 18 and 13 based on Danish first trimester data from 2005 to 2014. Prenat Diagn 2016;36:643–9.
8. Morris JK, Mutton D, Alberman E. Revised estimates of the maternal age specific live birth prevalence of Down's syndrome. J Med Screen 2002;9:2–6.
9. Balayla J, Azoulay L, Assayag J, Benjamin A, Abenhaim HA. Effect of maternal age on the risk of stillbirth: a population-based cohort study on 37 million births in the United States. Am J Perinatol 2011;28:643–50.
10. Waldenström U, Aasheim V, Nilsen AB, Rasmussen S, Pettersson HJ, Schytt E. Adverse pregnancy outcomes related to advanced maternal age compared with smoking and being overweight. Obstet Gynecol 2014;123:104–12.
11. Flenady V, Koopmans L, Middleton P, Frøen JF, Smith GC, Gibbons K, et al. Major risk factors for stillbirth in high-income countries: a systematic review and meta-analysis. Lancet 2011;377:1331–40.
12. Laopaiboon M, Lumbiganon P, Intarut N, Mori R, Ganchimeg T, Vogel JP, et al. Advanced maternal age and pregnancy outcomes: a multicountry assessment. BJOG 2014;121(suppl 1):49–56.
13. Kenny L, Lavender T, McNamee R, O'Neill SM, Mills T, Khashan AS. Advanced maternal age and adverse pregnancy outcome: evidence from a large contemporary cohort. PLoS One 2013;8:e56583.
14. Loane M, Dolk H, Morris JK; EUROCAT Working Group. Maternal age-specific risk of non-chromosomal anomalies. BJOG 2009;116:1111–9.
15. Goetzinger KR, Shanks AL, Odibo AO, Macones GA, Cahill AG. Advanced maternal age and the risk of major congenital anomalies. Am J Perinatol 2017;34:217–22.
16. Hollier LM, Leveno KJ, Kelly MA, Mcintire DD, Cunningham FG. Maternal age and malformations in singleton births. Obstet Gynecol 2000;96:701–6.
17. Gill SK, Broussard C, Devine O, Green RF, Rasmussen SA, Reefhuis J, et al. Association between maternal age and birth defects of unknown etiology: United States, 1997–2007. Birth Defects Res A Clin Mol Teratol 2012;94:1010–8.
18. Ekelund CK, Jorgensen FS, Petersen OB, Sundberg K, Tabor A. Impact of a new national screening policy for Down's syndrome in Denmark: population based cohort study. BMJ 2008;337:a2547.
19. Ekelund CK, Kopp TI, Tabor A, Petersen OB. The Danish Fetal Medicine database. Clin Epidemiol 2016;8:479–83.
20. European surveillance of congenital anomalies. Minor anomalies for exclusion. 2005. Available at: http://www.eurocat-network.eu/aboutus/datacollection/guidelinesforregistration/malformationcodingguides
. Retrieved February 18, 2017.
21. Catalano PM, Shankar K. Obesity and pregnancy: mechanisms of short term and long term adverse consequences for mother and child. BMJ 2017;357:j1.
22. Wallace ME, Mendola P, Kim SS, Epps N, Chen Z, Smarr M, et al. Racial/ethnic differences in preterm perinatal outcomes. Am J Obstet Gynecol 2017;216:306.e1–12.
23. De Souza E, Halliday J, Chan A, Bower C, Morris JK. Recurrence risks for trisomies 13, 18, and 21. Am J Med Genet A 2009;149A:2716–22.
24. Cnattingius S. The epidemiology of smoking during pregnancy: smoking prevalence, maternal characteristics, and pregnancy outcomes. Nicotine Tob Res 2004;6(suppl 2):S125–40.
25. Bai J, Wong FW, Bauman A, Mohsin M. Parity and pregnancy outcomes. Am J Obstet Gynecol 2002;186:274–8.
26. Marino JL, Moore VM, Willson KJ, Rumbold A, Whitrow MJ, Giles LC, et al. Perinatal outcomes by mode of assisted conception and sub-fertility in an Australian data linkage cohort. PLoS One 2014;9:e80398.
27. Farhangniya M, Dortaj Rabori E, Mozafari Kermani R, Haghdoost AA, Bahrampour A, Bagheri P, et al. Comparison of congenital abnormalities of infants conceived by assisted reproductive techniques versus infants with natural conception in Tehran. Int J Fertil Steril 2013;7:217–24.
28. Rothman KJ. Using regression models in epidemiological analysis. In: Epidemiology—an introduction. Oxford (United Kingdom): Oxford University Press; 2002. p. 181–98.
29. European surveillance of congenital anomalies. EUROCAT subgroups of congenital anomalies. 2014. Available at: http://www.eurocat-network.eu/aboutus/datacollection/guidelinesforregistration/malformationcodingguides
. Retrieved February 18, 2017.