Abstract
Background:
Fetal aneuploidy risk increases with maternal age, but the majority of pregnancies complicated by trisomy 21 occur in younger women. It has been suggested that grandmaternal and/or paternal age may also play a role.
Objectives:
To assess the association between grandmaternal and paternal age and trisomy 21.
Methods:
For the grandmaternal assessments, we included all offspring with trisomy 21 in a statewide birth defects surveillance system (1995–2015) that could be linked to 3-generation matrilineal pedigrees in the Utah Population Database. Ten sex/birth year-matched controls were selected for each case (770 cases and 7700 controls). For the paternal assessments, our cohort included all trisomy 21 cases (1995–2015) where both the mother and father resided in Utah at the time of birth (1409 cases and 14 090 controls). Ages were categorized by 5-year intervals (reference: 25–29 years). Conditional logistic regression, adjusting for potential confounding factors, was used to model the association between grandmaternal and paternal age and trisomy 21.
Results:
No association between grandmaternal age and trisomy 21 was detected, whether age was assessed continuously (adjusted odds ratio [OR] 1.01, 95% confidence interval [CI] 0.98, 1.03) or categorically after adjusting for grandmaternal and grandpaternal race/ethnicity and grandpaternal age. Compared to fathers aged 20–29 years, fathers <20 years (aOR 3.15, 95% CI 1.99, 4.98) and 20–24 years (aOR 1.39, 95% CI 1.11, 1.73) had increased odds of trisomy 21 offspring, after adjusting for maternal and paternal race/ethnicity and maternal age. Results were consistent after excluding stillbirths, multiples, and trisomy 21 due to translocation or mosaicism.
Conclusions:
Maternal age is an important risk factor for trisomy 21 offspring; however, this population-based study shows that that young paternal age is also associated with trisomy 21, after taking into account maternal age and race/ethnicity.
Keywords: case-control study, down syndrome, grandmaternal age, maternal age, paternal age, Trisomy 21
1 |. BACKGROUND
Trisomy 21, a chromosomal abnormality originating in most cases (95%) from three copies of chromosome 21, is one of the most common genetic causes of intellectual and developmental disabilities, affecting 1:691 livebirths in the United States.1 The association between advanced maternal age and trisomy 21 was first reported in 1933,2 and is now recognized as the major risk factor for having a birth with trisomy 21.3 While the strong positive maternal-age association has been well established, prior research supports a more complex scenario, whereby dietary, life style, environmental, occupational, genetic, and epigenetic factors may also play a role, not only from the mother, but also the father or maternal grandmother.3
One hypothesis is that the age of the mother’s mother (the fetuses’ maternal grandmother) at the time she delivered the fetuses’ mother may be an independent risk factor for the birth of a child with trisomy 21, perhaps via maternal gonadal trisomy 21 mosaicism4 as a result of her own mother’s age. It has also been hypothesized that the hormonal milieu of older mothers might influence ovum development in their female fetuses, thus increasing a daughter’s risk of having a pregnancy complicated by trisomy 21.5 More recently, it has been suggested that this putative transgenerational influence might be mediated by defective spindle-associated proteins6 or other epigenetic-mediated changes.7,8 Several evaluations of this hypothesis have been reported, some that confirmed this hypothesis6,9–11 and others that have not.12–14
It has been further hypothesized that advanced paternal age may contribute to increased risk of trisomy 21 among offspring in up to 21% of cases15 and up to 50% for parents over 40.16 Over a dozen studies have been conducted addressing this question, with four studies indicating older paternal age to be associated with higher odds of offspring with trisomy 2117–20 six studies finding younger age to be an important factor20–25 and six studies reporting a null relationship.26–31 In contrast to the hypothesized association between grandmaternal age and trisomy 21 in grandmaternal offspring; no purported risk has been hypothesized for grandpaternal age on trisomy 21 in grandpaternal offspring; with only one prior study assessing, with null findings.31
The inconsistent associations between grandmaternal and paternal age and trisomy 21 may be due in part to cross-sectional and/or underpowered study designs or inadequate control for important confounding or modifying factors. Indeed, beyond age, prior research has reported that Hispanic ethnicity increases trisomy 21 risk.32
We designed a population-based matched case-control study to examine the associations of trisomy 21 risk in the offspring in relation to maternal age, grandmaternal (fetuses’ mother’s mother) age, and paternal age. Given the increasing prevalence of trisomy 21 worldwide over the last decade,33,34 and growing number of births to women aged over 35,35 improved risk prediction and counselling might improve public health for current and future generations.
2 |. METHODS
2.1 |. Study population
This analysis relies on two population-based data resources. The first is the Utah Birth Defect Network (UBDN). In existence since 1994, the UBDN is a comprehensive statewide surveillance programme that identifies all major structural and chromosomal malformations occurring among resident women. The second key data source is the Utah Population Database (UPDB). The UPDB is a comprehensive, population-based genealogical and medical data resource, including linked birth and death records and UBDN records for approved projects. Originally, the UPDB was a set of electronic genealogies created in the late 1970s for 1.6 million individuals who were linked to create multigenerational pedigrees.36 Over time, the UPDB has expanded these genealogies (primarily through the use of vital records) and the number and types of records used. It currently includes information on over 11 million individuals, spanning 18 generations.37
2.2 |. Case-control selection
Of particular interest to this study are 3-generation triads represented in the multigenerational pedigrees within UPDB. Among the total pregnancies in Utah between 1995 and 2015, for whom the grandmother gave birth to the mother in Utah, 770 trisomy 21 cases were identified. Trisomy 21 pregnancies in the UBDN from 1995–2015 were linked to the UPDB and trisomy 21 cases who were grandchildren in the three-generation matrilineal pedigrees were identified. Ten sex and birth year-matched control infants (ie controls are those with no evidence of trisomy 21 themselves or their siblings) per case whose mother and maternal grandmother also resided in Utah at time of birth were also identified in the UPDB, resulting in 7700 controls. For the paternal age and trisomy 21 risk assessments, we used all trisomy 21 pregnancies in Utah between 1995 and 2015, for whom both the mother and father resided in Utah at the time of birth, resulting in 1409 cases. Ten sex and birth year-matched control infants per case whose mother and father also resided in Utah at the time of birth were identified in the UPDB, resulting in 14 090 controls. For cases resulting in a stillbirth (n = 53), termination (n = 6), or spontaneous abortion (n = 1) we used year on the fetal death certificate or birth defect network record.
2.3 |. Exposure
Paternal age was defined as his age at the date of index offspring’s birth as recorded in the birth certificate. Grandmaternal age was defined as her age at the date she delivered the index offspring’s mother. Age was assessed continuously and by 5-year intervals: <20, 20–24, 25–29 [reference], 30–34, 35–39, ≥40 years.
2.4 |. Outcome
Potential trisomy 21 cases were identified by the UBDN via multiple prenatal and postnatal sources. Each potential case was reviewed by a clinical geneticist to confirm that the trisomy 21 case met eligibility criteria. With appropriate IRB approvals, UBDN also provided us with karyotype of each case: nondisjunction trisomy, mosaic trisomy, translocation trisomy, or unknown trisomy.
2.5 |. Statistical analysis
To examine the association between maternal age, grandmaternal age, paternal age, and trisomy 21, we used conditional logistic regression, accounting for matching of cases to controls, to calculate odds ratios (OR) and 95% confidence intervals (CI). We selected covariates separately for each model, identifying confounders as factors that might affect both the exposure (maternal/grandmaternal/paternal age) and outcome of trisomy 21 including appropriate kin ages and race/ethnicity. Effect modification by age (continuous) and race/ethnicity were also examined. The full analyses include all pregnancy outcomes, including spontaneous abortions, stillbirths, pregnancy terminations, and multiple births and all trisomy 21 types, to be comparable to prior studies. All analyses were conducted using SAS 9.4.
2.5.1 |. Sensitivity analysis
We performed three sensitivity analyses restricting the population to only (a) livebirths; (b) singleton births; or (c) trisomy 21 due to nondisjunction. Additionally, to address unmeasured confounding such as dietary, life style, environmental, or occupational factors that may impact the association between grandmaternal or paternal age and trisomy 21, we used the E-value method.38
2.5.2 |. Missing data
To address potential missing data bias, we performed multiple imputation using multivariate normal distribution (with 50 imputations) for our adjusted conditional logistic regression models. For the grandmaternal and maternal analyses with 770 trisomy 21 cases and 7700 controls, missing data for covariates was n = 13 (0.2%) for maternal race, n = 8 (0.1%) for maternal ethnicity, n = 27 (0.3%) for grandmaternal race, n = 644 (7.6%) for grandmaternal ethnicity, n = 567 (6.7%) for paternal age, n = 593 (7%) for paternal race; n = 154 (1.8%) for paternal ethnicity, n = 511 (6.0%) for grandpaternal age, n = 942 (11.1%) for grandpaternal ethnicity and n = 478 (5.6%) for grandpaternal race. For the paternal analyses with 1409 trisomy 21 cases and 14 090 controls, missing data for covariates was n = 0 (0%) for maternal age, n = 55 (0.4%) for paternal race, n = 197 (1.3%) for paternal ethnicity, n = 7 (0.1%) for maternal race, and n = 98 (0.6%) for maternal ethnicity.
2.6 |. Ethics approval
Institutional Review Board approval was obtained from the Utah Department of Health and University of Utah.
3 |. RESULTS
Among the maternal/grandmaternal population, offspring with trisomy 21 (n = 770) and control offspring (n = 7700), mean maternal age was 27.0 (standard deviation [SD] 5.6; range: 13–48 years). Mean grandmaternal and paternal age at the time of maternal delivery was 25.8 (SD 5.6; range: 12–46 years) and 29.6 (SD 6.1; range 15–66), respectively. Maternal, grandmaternal, and paternal race was predominately white and non-Hispanic (Table 1). Among the maternal/grandmaternal trisomy 21 population (n = 770 cases), 704 (91.4%) had nondisjunction trisomy, 15 (2.0%) mosaic trisomy, 27 (3.5%) translocation trisomy, and 24 (3.1%) unknown trisomy 21 karyotype. Among the maternal/paternal trisomy 21 population (n = 1409 cases), 1303 (92.5%) had nondisjunction trisomy, 23 (1.6%) mosaic trisomy, 45 (3.2%) translocation trisomy, and 38 (2.7%) unknown trisomy 21 karyotype.
TABLE 1.
Population characteristics, 3-generation matrilineal pedigrees, 1995–2015
| Characteristics | Overall | Trisomy 21 offspring (n = 770) | Control offspring (n = 7700) |
|---|---|---|---|
| Maternal age (years)a | 27.0 (5.6) | 30.5 (7.1) | 26.7 (5.3) |
| Maternal race, n (%) | |||
| White | 7978 (94.3) | 733 (95.2) | 7245 (94.3) |
| Non-White | 479 (5.7) | 37 (4.8) | 442 (5.8) |
| Maternal ethnicity, n (%) | |||
| Hispanic | 995 (11.8) | 80 (10.4) | 915 (11.9) |
| Non-hispanic | 7467 (88.2) | 688 (89.6) | 6779 (88.1) |
| Grandmaternal agea (years) | 25.8 (5.6) | 26.2 (5.9) | 25.7 (5.5) |
| Grandmaternal race, n (%) | |||
| White | 8117 (96.1) | 744 (97.4) | 7373 (96.0) |
| Non-White | 326 (3.9) | 20 (2.6) | 306 (4.0) |
| Grandmaternal ethnicity, n (%) | |||
| Hispanic | 625 (8.0) | 47 (6.7) | 578 (8.1) |
| Non-Hispanic | 7201 (92.0) | 655 (93.3) | 6546 (91.9) |
| Paternal agea (years) | 29.6 (6.1) | 32.9 (7.5) | 29.2 (5.8) |
| Paternal race, n (%) | |||
| White | 7393 (93.9) | 691 (94.5) | 6702 (93.8) |
| Non-White | 484 (6.1) | 40 (5.5) | 444 (6.2) |
| Paternal ethnicity, n (%) | |||
| Hispanic | 778 (9.4) | 65 (8.6) | 713 (9.4) |
| Non-Hispanic | 7538 (90.6) | 689 (91.4) | 6849 (90.6) |
| Grandpaternal agea (years) | 28.5 (6.4) | 28.9 (6.7) | 28.4 (6.4) |
| Grandpaternal race, n (%) | |||
| White | 7722 (96.6) | 670 (98.0) | 7052 (96.5) |
| Non-White | 270 (3.4) | 14 (2.1) | 256 (3.5) |
| Grandpaternal ethnicity, n (%) | |||
| Hispanic | 534 (7.1) | 43 (6.8) | 491 (7.1) |
| Non-Hispanic | 6994 (92.9) | 590 (93.2) | 6404 (92.9) |
| Offspring sex, n (%) | |||
| Male | 4587 (54.2) | 417 (54.2) | 4170 (54.2) |
| Female | 3883 (45.8) | 353 (45.8) | 3530 (45.8) |
| Multiplicity | |||
| Singleton | 8191 (96.7) | 752 (97.6) | 7439 (96.6) |
| Twin | 266 (3.1) | 18 (2.3) | 248 (3.2) |
| Triplet | 13 (0.2) | 0 (0) | 13 (0.2) |
Note: Missing n = 0 for maternal age and grandmaternal age, n = 13 (0.2%) for maternal race, n = 8 (0.1%) for maternal ethnicity, n = 27 (0.3%) for grandmaternal race, n = 644 (7.6%) for grandmaternal ethnicity, n = 567 (6.7%) for paternal age, n = 593 (7%) for paternal race; n = 154 (1.8%) for paternal ethnicity, n = 511 (6.0%) for grandpaternal age, n = 942 (11.1%) for grandpaternal ethnicity and n = 478 (5.6%) for grandpaternal race.
Data are reported as mean (standard deviation).
Regarding the association between maternal age and trisomy 21, the adjusted OR for each additional year of maternal age was 1.10 (95% CI 1.07, 1.13; Table 2). When assessed categorically, women 35–39 had a near fourfold higher adjusted odds and women ≥40 years had over a twelve-fold higher odds of having offspring with trisomy 21 compared to women 25–29 years. No association between grandmaternal age and trisomy 21 was detected, whether age was assessed continuously or categorically.
TABLE 2.
Odds ratios (95% CI) for trisomy 21 by maternal, grandmaternal, and paternal age
| n (%) | Odds ratio (95% confidence interval) | |||
|---|---|---|---|---|
| Trisomy 21 offspring | Control offspring | Unadjusted | Adjusteda | |
| Maternal age (years) | n = 770 | n = 7700 | ||
| Continuous | 1.12 (1.11, 1.14) | 1.10 (1.07, 1.13) | ||
| <20 | 40 (5.2) | 602 (7.8) | 0.98 (0.69, 1.40) | 1.63 (1.01, 2.66) |
| 20–24 | 148 (19.2) | 2226 (28.9) | 0.99 (0.78, 1.24) | 1.15 (0.87, 1.52) |
| 25–29 | 178 (23.1) | 2677 (33.8) | 1.00 (Reference) | 1.00 (Reference) |
| 30–34 | 153 (19.9) | 1590 (20.6) | 1.45 (1.16, 1.82) | 1.27 (0.96, 1.68) |
| 35–39 | 157 (20.4) | 515 (6.7) | 4.74 (3.73, 6.02) | 3.47 (2.44, 4.95) |
| ≥40 | 94 (12.2) | 90 (1.2) | 15.63 (11.24, 21.74) | 12.35 (7.14, 21.73) |
| Grandmaternal age (years) | n = 770 | n = 7700 | ||
| Continuous | 1.02 (1.00, 1.03) | 1.01 (0.98, 1.03) | ||
| <20 | 83 (10.8) | 893 (11.6) | 0.89 (0.68, 1.15) | 0.98 (0.68, 1.40) |
| 20–24 | 249 (32.3) | 2717 (35.3) | 0.87 (0.72, 1.06) | 0.89 (0.70, 1.12) |
| 25–29 | 232 (30.1) | 2214 (28.8) | 1.00 (Reference) | 1.00 (Reference) |
| 30–34 | 126 (16.4) | 1280 (16.3) | 0.94 (0.75, 1.18) | 0.90 (0.68, 1.20) |
| 35–39 | 61 (7.9) | 491 (6.4) | 1.19 (0.88, 1.60) | 1.31 (0.88, 1.94) |
| ≥40 | 19 (2.5) | 104 (1.4) | 1.73 (1.05, 2.87) | 1.76 (0.89, 3.46) |
| Paternal age (years) | n = 1409 | n = 14 090 | ||
| Continuous | 1.09 (1.08, 1.10) | 1.00 (0.98, 1.02) | ||
| <20 | 25 (1.8) | 317 (2.3) | 1.23 (0.80, 1.89) | 3.15 (1.99, 4.98) |
| 20–24 | 132 (9.4) | 2459 (17.5) | 0.84 (0.68, 1.04) | 1.39 (1.11, 1.73) |
| 25–29 | 305 (21.7) | 4821 (34.2) | 1.00 (Reference) | 1.00 (Reference) |
| 30–34 | 327 (23.2) | 3823 (27.1) | 1.36 (1.16, 1.60) | 0.77 (0.64, 0.92) |
| 35–39 | 318 (22.6) | 1779 (12.6) | 2.83 (2.40, 3.35) | 0.97 (0.79, 1.20) |
| ≥40 | 302 (21.4) | 891 (6.3) | 5.44 (4.56, 6.48) | 1.20 (0.94, 1.54) |
For maternal age model: Adjusted for maternal race/ethnicity, paternal age and paternal race/ethnicity, grandmaternal/grandpaternal age and grandmaternal/grandpaternal race/ethnicity. For grandmaternal age model: Adjusted for grandmaternal/grandpaternal race/ethnicity and grandpaternal age. For paternal age model: Adjusted for maternal and paternal race/ethnicity and maternal age.
We found no associations between continuous paternal age and offspring trisomy 21. When assessed categorically, fathers <20 years had over a threefold higher odds those 20–24 years had a 1.4 higher odds of having offspring with trisomy 21 compared to men 25–29 years.
While we found no effect modification by grandmother’s age in the relationship between maternal age and trisomy 21, we did find paternal age to moderate the relationship between maternal age and trisomy 21 (Pinteraction < 0.001). Stratified analyses examining the odds of trisomy 21 by maternal and grandmaternal age categories confirm the lack of appreciable effect modification of grandmaternal age in the relationship between maternal age and trisomy 21, both in the unadjusted (Figure 1) and grandmaternal and maternal race and ethnicity adjusted models (Table S1). In contrast, there is some indication (albeit less precise) that young paternal age in combination with old maternal age, a relative uncommon combination,39 has the highest association with trisomy 21 in offspring. Fathers <24 partnered with mothers ≥35 years had a near eighteen-fold higher odds of trisomy 21 compared to fathers 25–29 partnered with mothers 25–29 years (Figure 2), and trending in the same direction for fathers <24 partnered with mothers 30–34 years. This pattern held true after adjustment for paternal and maternal race and ethnicity (Table S2). No effect modification by race or ethnicity was found for any of the groups examined.
FIGURE 1.
Unadjusted odds ratios (95% CI) for grandmaternal and maternal combined age categories and trisomy 21
FIGURE 2.
Unadjusted odds ratios (95% CI) for paternal and maternal combined age categories and trisomy 21
3.1 |. Sensitivity analysis
The associations remained robust after restricting analyses to only women who had a livebirth, singletons, or trisomy 21 due to nondisjunction (Table 3), or after addressing potential missing data bias via multiple imputation (Table 4). Regarding unmeasured confounding, the observed odds ratio of 3.2 for fathers <20 years and 1.4 for fathers 20–24 could be explained away by an unmeasured confounder that was associated with both the exposure and the outcome by an odds ratio of 5.8-fold and 3.4-fold, respectively, above and beyond the measured confounders, but weaker confounding could not do so. The confidence interval could be moved to include the null by an unmeasured confounder that was associated with both the exposure and the outcome by an odds ratio of 2.1-fold and 1.5-fold, respectively, above and beyond the measured confounders, but weaker confounding could not do so.
TABLE 3.
Adjusted odds ratio (95% confidence interval) for trisomy 21 by maternal, grandmaternal, and paternal age; restricted to only livebirths, only singletons, only nondisjunction type
| Livebirths | Singletons | Nondisjunction | |
|---|---|---|---|
| Adjusteda odds ratio (95% confidence interval) | |||
| Maternal age (years) | |||
| Continuous | 1.09 (1.06, 1.12) | 1.10 (1.07, 1.13) | 1.10 (1.07, 1.14) |
| <20 | 1.77 (1.08, 2.88) | 1.57 (0.96, 2.56) | 1.65 (0.98, 2.77) |
| 20–24 | 1.23 (0.93, 1.63) | 1.12 (0.85, 1.49) | 1.21 (0.90, 1.62) |
| 25–29 | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) |
| 30–34 | 1.36 (1.02, 1.81) | 1.27 (0.95, 1.68) | 1.36 (1.01, 1.82) |
| 35–39 | 3.52 (2.45, 5.05) | 3.56 (2.48, 5.11) | 3.76 (2.59, 5.46) |
| ≥40 | 12.52 (7.18, 21.82) | 14.08 (7.94, 24.96) | 13.40 (7.57, 23.71) |
| Grandmaternal age (years) | |||
| Continuous | 1.01 (0.98, 1.03) | 1.00 (0.97, 1.03) | 1.01 (0.98, 1.04) |
| <20 | 0.96 (0.66, 1.38) | 1.03 (0.72, 1.48) | 0.87 (0.59, 1.28) |
| 20–24 | 0.92 (0.72, 1.17) | 0.90 (0.71, 1.15) | 0.93 (0.73, 1.19) |
| 25–29 | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) |
| 30–34 | 1.37 (0.91, 2.05) | 0.92 (0.69, 1.23) | 0.88 (0.65, 1.19) |
| 35–39 | 1.86 (0.94, 3.70) | 1.29 (0.86, 1.93) | 1.31 (0.86, 1.98) |
| ≥40 | 1.86 (0.94, 3.70) | 1.55 (0.76, 3.17) | 2.11 (1.04, 4.25) |
| Paternal age (years) | |||
| Continuous | 1.00 (0.99, 1.02) | 1.00 (0.99, 1.01) | 1.00 (0.98, 1.01) |
| <20 | 3.22 (2.03, 5.09) | 3.19 (2.01, 5.05) | 3.43 (2.15, 5.50) |
| 20–24 | 1.41 (1.13, 1.77) | 1.38 (1.10, 1.73) | 1.47 (1.17, 1.85) |
| 25–29 | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) |
| 30–34 | 0.77 (0.64, 0.92) | 0.75 (0.62, 0.89) | 0.78 (0.65, 0.94) |
| 35–39 | 0.97 (0.79, 1.20) | 0.95 (0.77, 1.17) | 0.97 (0.79, 1.20) |
| ≥40 | 1.21 (0.94, 1.56) | 1.16 (0.90, 1.50) | 1.18 (0.92, 1.53) |
For maternal age model: Adjusted for maternal race/ethnicity, paternal age and paternal race/ethnicity, grandmaternal/grandpaternal age and grandmaternal/grandpaternal race/ethnicity. For grandmaternal age model: Adjusted for grandmaternal/grandpaternal race/ethnicity and grandpaternal age. For paternal age model: Adjusted for maternal race/ethnicity and maternal age.
TABLE 4.
Odds ratios (95% CI) for trisomy 21 by maternal, grandmaternal, and paternal age using multiple imputation to address potential missing data biasa
| Trisomy 21 offspring | Control offspring | Adjusted OR (95% CI)b | |
|---|---|---|---|
| n (%) | n (%) | ||
| Maternal age | n = 770 | n = 7700 | |
| Continuous | 1.10 (1.08, 1.13) | ||
| <20 | 40 (5.2) | 602 (7.8) | 1.31 (0.88, 1.94) |
| 20–24 | 148 (19.2) | 2226 (28.9) | 1.12 (0.88, 1.43) |
| 25–29 | 178 (23.1) | 2677 (33.8) | 1.00 (Reference) |
| 30–34 | 153 (19.9) | 1590 (20.6) | 1.26 (0.99, 1.61) |
| 35–39 | 157 (20.4) | 515 (6.7) | 3.61 (2.69, 4.84) |
| ≥40 | 94 (12.2) | 90 (1.2) | 10.20 (6.68, 15.59) |
| Grandmaternal age | n = 770 | n = 7700 | |
| Continuous | 1.01 (0.98, 1.03) | ||
| <20 | 83 (10.8) | 893 (11.6) | 0.92 (0.67, 1.26) |
| 20–24 | 249 (32.3) | 2717 (35.3) | 0.88 (0.72, 1.08) |
| 25–29 | 232 (30.1) | 2214 (28.8) | 1.00 (Reference) |
| 30–34 | 126 (16.4) | 1280 (16.3) | 0.93 (0.73, 1.20) |
| 35–39 | 61 (7.9) | 491 (6.4) | 1.17 (0.81, 1.67) |
| ≥40 | 19 (2.5) | 104 (1.4) | 1.70 (0.94, 3.10) |
| Paternal Age | n = 1409 | n = 14 090 | |
| Continuous | 1.00 (0.99, 1.01) | ||
| <20 | 25 (1.8) | 317 (2.3) | 3.30 (2.10, 5.17) |
| 20–24 | 132 (9.4) | 2459 (17.5) | 1.38 (1.11, 1.72) |
| 25–29 | 305 (21.7) | 4821 (34.2) | 1.00 (Reference) |
| 30–34 | 327 (23.2) | 3823 (27.1) | 0.78 (0.65, 0.92) |
| 35–39 | 318 (22.6) | 1779 (12.6) | 0.97 (0.79, 1.19) |
| ≥40 | 302 (21.4) | 891 (6.3) | 1.19 (0.93, 1.52) |
For maternal and grandmaternal age data set: Missing n = 0 for maternal age and grandmaternal age, n = 13 (0.2%) for maternal race, n = 8 (0.1%) for maternal ethnicity, n = 27 (0.3%) for grandmaternal race, n = 644 (7.6%) for grandmaternal ethnicity, n = 567 (6.7%) for paternal age, n = 593 (7%) for paternal race; n = 154 (1.8%) for paternal ethnicity, n = 511 (6.0%) for grandpaternal age, n = 942 (11.1%) for grandpaternal ethnicity and n = 478 (5.6%) for grandpaternal race. For paternal age data set: Missing n = 0 for maternal age and paternal age, n = 7 (0.1%) for maternal race, n = 98 (0.6%) for maternal ethnicity, n = 55 (0.4%) for paternal race; n = 197 (1.3%) for paternal ethnicity.
For maternal age model: Adjusted for maternal race/ethnicity, paternal age and paternal race/ethnicity, grandmaternal/grandpaternal age and grandmaternal/grandpaternal race/ethnicity. For grandmaternal age model: Adjusted for grandmaternal/grandpaternal race/ethnicity and grandpaternal age. For paternal age model: Adjusted for maternal and paternal race/ethnicity and maternal age.
4 |. COMMENT
4.1 |. Principal findings
In this population-based study we found that maternal, but not grandmaternal, age was associated with the probability of trisomy 21, with a nearly 3.5-fold higher adjusted odds for mothers ages 35 to 39 and a nearly 12.5-fold higher odds for mothers over age 40, compared to mothers ages 25–29. While we found no association between grandmaternal age and probability of trisomy 21 in matrilineal grandchildren, we found that fathers aged <20 had a 3.2-fold higher odds and fathers 20–24 years of age had a 1.4-fold higher adjusted odds of having a trisomy 21 offspring, compared to fathers age 25–29.
4.2 |. Strengths of the study
Our study has several strengths. It is population-based and thus less likely than previous reports to suffer from referral or access biases. Because of the de-identified nature of the data set and the population coverage of the relevant databases, all eligible cases were included, thus limiting selection or recruitment biases. Also, the cases and controls were born from the same populations in the same time frames, thus limiting the effects of any changes in social patterns. Records were also specifically reviewed to ensure that no genetic misclassifications due to adoptions were included. In addition, all trisomy 21 cases were reviewed by expert clinicians to confirm the diagnosis. Perhaps most importantly, we were able to identify additional cases of trisomy 21 from stillbirth and selected pregnancy termination records. Pregnancies complicated by trisomy 21 are known to be at increased risk for spontaneous miscarriage40 and stillbirth41 and our ability to include at least a portion of these cases lends additional validity to our conclusions, despite our inability to capture losses among our control population.
4.3 |. Limitations of the data
Our study also has some limitations. While our data does include a limited number of stillbirths, terminations, and spontaneous abortions (~4%), an improvement over many of the prior studies that exclude all non-livebirths, we are unable to include those trisomy 21 cases that resulted in an early miscarriage but were not confirmed by karyotype. An estimated 67%−74% of trisomy 21 pregnancies in the US end with a spontaneous or therapeutic termination. Data limitations did not allow us to analyse all these cases. As a result, our findings may be biased if the parental and grandparental characteristics of observed trisomy 21 cases among livebirths differ fundamentally from the characteristics of trisomy 21 cases that ended before a livebirth.42 Privacy issues in regards to spontaneous or planned terminations make access to such data difficult to obtain; however, future studies that can include early karyotyped losses (planned and spontaneous) in addition to stillbirths are needed to rule out selection bias in future studies.
The increased ability to prenatally diagnose trisomy 21 over the past several decades could have also introduced a selection bias towards younger trisomy 21 mothers and maternal grandmothers. While our study was one of the largest studies to date, we were limited in our ability to assess effect modification by age, notably in regards to paternal and maternal subgroupings (few counts for young fathers partnered with old mothers). Furthermore, we did not have information on whether nondisjunction meiotic errors were of paternal versus maternal origin. Additionally, we are unable to address the possibility of life style or environmental influences, although sensitivity analyses for unmeasured confounding provided reassurance.
4.4 |. Interpretation
4.4.1 |. Maternal age and trisomy 21
Our finding of higher trisomy 21 probability with advanced maternal age >35 was expected, given prior evidence.2,3 What is novel about our study is our ability to contribute to the knowledge gap as to whether Hispanic ethnicity contributes to trisomy 21 risk.32 A prior study using National Center for Health Statistics Data (1989 to 1991) reported an unadjusted OR of 3.2 (95% CI 3.0, 3.5) for trisomy 21 among non-Hispanic white mothers of ≥35 vs <35 years, compared to an OR of 6.5 (95% CI 5.4, 7.7) among Mexican Americans ages ≥35 vs <35 years.32 We, however, found no evidence for effect modification by race/ethnicity in the relationship between maternal age and trisomy 21 in offspring. Given the different time frames and geographical areas between our and the Khoshnood et al study,32 more research is needed regarding the interplay between maternal age, ethnicity, and trisomy 21 risk.
4.4.2 |. Grandmaternal age and trisomy 21
Although several previous studies have reported an association between grandmaternal age and subsequent trisomy 21 rates in their grandchildren, we found no evidence for such an association. The majority of the represented populations from prior studies has been from referral clinics, and may therefore have been biased by socio-economic factors known to be associated with increased access to health care.43 In this particular situation, older grandparents may have accumulated more wealth and life-experience and thus have been more likely to encourage and/or facilitate evaluations of their grandchildren with trisomy 21 by referral specialists. Additionally, prior studies looking at the association between grandmaternal age and trisomy 21 risk are limited by overlooking important confounding factors or from adjustment bias by controlling for downstream variables (ie maternal age when assessing impact of grandmaternal age on trisomy 21 risk).
Our null finding between grandmaternal age and odds of grand-offspring trisomy 21 is consistent with the one prior population-based study,12 which analysed data from two population-based, case-control studies: the Atlanta Down Syndrome Project (1989–1999) and the National Down Syndrome Project (2001–2004), enrolling 1215 case families and 2293 controls.
4.4.3 |. Paternal age and trisomy 21
Our finding of an inverse association between paternal age and offspring with trisomy 21 is in agreement with several prior studies,20–25 including a large population-based Swiss study of nearly two million births evaluated between 1979 and 2006.21 In that study, Steiner et al21 found that younger fathers have the higher probability of having a trisomy 21 child compared to older fathers after accounting for maternal age. Our findings are also in agreement with a recent US study using data from approximately 12 million birth records (2014–2016), which showed that increasing paternal age is associated with a decrease in the birth prevalence of trisomy 21.22 This inverse paternal-age association reporting dates back to 1983, within an Ohio case-control study among 1244 trisomy 21 cases,24 and noted again in 1995 in a British Columbia case-control study among 997 trisomy 21 cases,20 both of which adjusted for maternal age. Two prior studies have assessed the interactive effects of maternal and paternal age on trisomy 21 risk,23,26 with one reporting that young fathers had a higher risk of trisomy 21 in offspring when the mothers were older but not younger,23 similar to our study, while the other study reporting no association between paternal age and trisomy 21 nor effect modification.26
One concern regarding our and others’ observed inverse paternal-age effect is the potential for selection bias. A prior study25 reported a 1.25 (95% CI 0.93, 1.65) and 1.28 (95% CI 1.08, 1.51) higher adjusted prevalence ratio for trisomy 21 among fathers <20 and 20–24 years, respectively (compared to fathers age 25–29 years). However, the study excluded missing fathers in 20% of trisomy 21 cases from the analysis, likely introducing selection bias, as the authors concluded.25 The current study contained no missing paternal data in 1409 trisomy 21 cases, so selection bias is unlikely due to this cause.
The mechanisms underlying paternal chromosomal nondisjunction is poorly understood.21 It is widely reported that trisomy 21 nondisjunctional error is >95% maternal and ≤5% paternal.44 However, prior research has revealed that the paternal contribution may be as high as 21%.15 Whether younger, rather than older, men may be more prone to nondisjunction of chromosome 21 occurring during spermatogenesis is not known. As noted previously, different environmental, occupational, and life style risk factors among younger fathers may be at play.20 An alternative explanation is that fertility of a couple is higher the younger the father is in couples with advanced maternal age, and therefore an increased risk for children with aneuploidies is not unexpected.21
In regards to association between older paternal age and trisomy 21 offspring, we found nearly a three- and sixfold higher unadjusted odds for fathers 35–39 and father ≥40 years, compared to fathers 25–29 years, but these associations disappeared after adjustments for maternal age and maternal/paternal race/ethnicity. These null findings are consistent with the majority of studies.21–31 Prior studies that have found a positive association between paternal age and trisomy 21 in offspring were limited in number of trisomy 21 cases included,18,19 found an increased risk only in fathers ≥50 vs <50 years,17 or interestingly found a U-shaped association with younger fathers having a twofold higher odds than older fathers.20
4.5 |. Conclusions
The predominant influence on trisomy 21 risk was maternal age at delivery in this population-based study. Additionally, while we found no association with grandmaternal age, we found younger paternal age to be associated with a higher odds of trisomy 21 in the offspring, suggesting that the mother’s age may not be the sole consideration when assessing a couple’s risk. While these findings, in agreement with six prior studies,20–25 may not impact clinical counselling among couples trying to conceive, if corroborated by further research, they may inform counselling in assisted reproductive technology.21
Supplementary Material
Synopsis.
Study question
Are paternal or grandmaternal age associated with higher probability of trisomy 21 (Down syndrome) in offspring?
What is already known
Advanced maternal age is a major risk factor for trisomy 21; however, the majority of babies with trisomy 21 are born to mothers less than 35 years suggesting a more complex scenario of potential risk factors. Grandmaternal and paternal age have been hypothesized to play a role, yet prior research is equivocal.
What this study adds
Findings from this study indicate that grandmaternal age is not associated with trisomy 21 in matrilineal grandchildren; however, young paternal age is associated with having offspring with trisomy 21, after taking into account maternal age and other potential confounding factors.
ACKNOWLEDGEMENTS
We thank the Pedigree and Population Resource of Huntsman Cancer Institute and the Utah Department of Health for ongoing collection, maintenance, and support of the UPDB and UBDN. We also acknowledge partial support through grant P30 CA2014 from NCI and University of Utah Personalized Health and Center for Clinical and Translational Science.
Funding information
This project is supported by the Health Resources and Services Administration (HRSA) of the US Department of Health and Human Services (HHS) under the “Maternal and Child Health Services” grant (B04MC25374). This information or content and conclusions are those of the authors and should not be construed as the official position or policy of, nor should any endorsements be inferred by HRSA, the US Government or the Utah Department of Health. Also supported in part by R01-AG022095 (KRS); K12-HD085852 (HH); K01-AGA58781 (KCS), and UL-TR001067 and HA and Edna Benning Presidential Endowment (MWV).
Footnotes
CONFLICT OF INTEREST
The authors declare no conflict of interest.
SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section.
REFERENCES
- 1.Patterson D Molecular genetic analysis of Down syndrome. Human Genet. 2009;126:195–214. [DOI] [PubMed] [Google Scholar]
- 2.Penrose LS. The relative effects of paternal and maternal age in mongolism. J Genet. 1933;27:219–224. [DOI] [PubMed] [Google Scholar]
- 3.Coppede F Risk factors for Down syndrome. Arch Toxicol. 2016;90:2917–2929. [DOI] [PubMed] [Google Scholar]
- 4.Penrose L Dermatoglyphics in mosaic mongolism and allied conditions. Genet Today. 1965;3:973. [Google Scholar]
- 5.Mikkelsen M Down anomaly: new research aspects of an old and well known syndrome. Prog Clin Biol Res. 1985;177:293. [PubMed] [Google Scholar]
- 6.Malini SS, Ramachandra NB. Influence of advanced age of maternal grandmothers on Down syndrome. BMC Med Genet. 2006;7:4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Bentov Y, Yavorska T, Esfandiari N, Jurisicova A, Casper RF. The contribution of mitochondrial function to reproductive aging. J Assist Reprod Genet. 2011;28:773–783. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ge Z-J, Schatten H, Zhang C-L, Sun Q-Y. Oocyte ageing and epigenetics. Reproduction. 2015;149:R103–R114. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Aagesen L, Grinsted J, Mikkelsen M. Advanced grandmaternal age on the mother’s side–a risk of giving rise to trisomy 21. Ann Hum Genet. 1984;48:297–301. [DOI] [PubMed] [Google Scholar]
- 10.Greenberg R Two factors influencing the births of Mongols to younger mothers. Med Officer. 1963;109:62–64. [Google Scholar]
- 11.Papp Z, Varadi E, Szabo Z. Grandmaternal age at birth of parents of children with trisomy 21. Hum Genet. 1977;39:221–224. [DOI] [PubMed] [Google Scholar]
- 12.Allen EG, Freeman SB, Druschel C, et al. Maternal age and risk for trisomy 21 assessed by the origin of chromosome nondisjunction: a report from the Atlanta and National Down Syndrome Projects. Hum Genet. 2009;125:41–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Forssman H, Akesson H. Consanguineous marriages and mongolism. CIBA Foundation Study Group; 1967. [Google Scholar]
- 14.Stoller A, Collmann R. Grandmaternal age at birth of mothers of children with Down’s syndrome (Mongolism). J Intellect Disab Res. 1969;13:201–205. [DOI] [PubMed] [Google Scholar]
- 15.Jyothy A, Kumar KS, Mallikarjuna GN, et al. Parental age and the origin of extra chromosome 21 in Down syndrome. J Hum Genet. 2001;46:347–350. [DOI] [PubMed] [Google Scholar]
- 16.Fisch H, Hyun G, Golden R, Hensle TW, Olsson CA, Liberson GL. The influence of paternal age on down syndrome. J Urol. 2003;169:2275–2278. [DOI] [PubMed] [Google Scholar]
- 17.Erickson JD, Bjerkedal TO. Down syndrome associated with father’s age in Norway. J Med Genet. 1981;18:22–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Stene J, Stene E, Stengel-Rutkowski S, Murken JD. Paternal age and Down’s syndrome: data from prenatal diagnoses (DFG). Hum Genet. 1981;59:119–224. [DOI] [PubMed] [Google Scholar]
- 19.Zhu JL, Made KM, Vestergaard M, Olesen AV, Basso O, Olsen J. Paternal age and congenital malformations. Hum Reprod. 2005;20:3173–3177. [DOI] [PubMed] [Google Scholar]
- 20.McIntosh GC, Olshan AF, Baird PA. Paternal age and the risk of birth defects in offspring. Epidemiology. 1995;6:282–288. [DOI] [PubMed] [Google Scholar]
- 21.Steiner B, Masood R, Rufibach K, et al. An unexpected finding: younger fathers have a higher risk for offspring with chromosomal aneuploidies. Eur J Hum Genet. 2015;23:466–472. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Thompson JA. Disentangling the roles of maternal and paternal age on birth prevalence of down syndrome and other chromosomal disorders using a Bayesian modeling approach. BMC Med Res Methodol. 2019;19:82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Agopian AJ, Marengo LK, Mitchell LE. Predictors of trisomy 21 in the offspring of older and younger women. Birth Defects Res A Clin Mol Teratol. 2012;94:31–35. [DOI] [PubMed] [Google Scholar]
- 24.Roecker GO, Huether CA. An analysis for paternal-age effect in Ohio’s Down syndrome births, 1970–1980. Am J Hum Genet. 1983;35:1297–1306. [PMC free article] [PubMed] [Google Scholar]
- 25.Archer NP, Langlois PH, Suarez L, Brender J, Shanmugam R. Association of paternal age with prevalence of selected birth defects. Birth Defects Res A Clin Mol Teratol. 2007;79:27–34. [DOI] [PubMed] [Google Scholar]
- 26.Lian ZH, Zack MM, Erickson JD. Paternal age and the occurrence of birth defects. Am J Hum Genet. 1986;39:648–660. [PMC free article] [PubMed] [Google Scholar]
- 27.De Souza E, Alberman E, Morris JK. Down syndrome and paternal age, a new analysis of case-control data collected in the 1960s. Am J Med Genet A. 1960s;149A:1205–1208. [DOI] [PubMed] [Google Scholar]
- 28.Cross PK, Hook EB. An analysis of paternal age and 47,+21 in 35,000 new prenatal cytogenetic diagnosis data from the New York State Chromosome Registry: no significant effect. Hum Genet. 1987;77:307–316. [DOI] [PubMed] [Google Scholar]
- 29.Yang Q, Wen SW, Leader A, Chen XK, Lipson J, Walker M. Paternal age and birth defects: how strong is the association? Hum Reprod. 2007;22:696–701. [DOI] [PubMed] [Google Scholar]
- 30.Oliver TR, Bhise A, Feingold E, Tinker S, Masse N, Sherman SL. Investigation of factors associated with paternal nondisjunction of chromosome 21. Am J Med Genet A. 2009;149A:1685–1690. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Kazaura MR, Lie RT, Skjaerven RS. Grandparents’ age and the risk of Down’s syndrome in Norway. Acta Gynecol Scand. 2006;85:236–240. [DOI] [PubMed] [Google Scholar]
- 32.Khoshnood B, Pryde P, Wall S, Singh J, Mittendorf R, Lee KS. Ethnic differences in the impact of advanced maternal age on birth prevalence of Down syndrome. Am J Public Health. 2000;90:1778–1781. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.St Louis AM, Kim K, Browne ML, et al. Prevalence trends of selected major birth defects: A multi-state population-based retrospective study, United States, 1999 to 2007. Birth Defects Res. 2017;109:1442–1450. [DOI] [PubMed] [Google Scholar]
- 34.Loane M, Morris JK, Addor MC, 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. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Hamilton BE, Martin JA, Osterman MJK, Rossen LM. Births: Provisional data for 2018. Vital Statistics Rapid Release; no 7. Hyattsville, MD: National Center for Health Statistics. May 2019. [PubMed] [Google Scholar]
- 36.Cannon Albright LA. Utah family-based analysis: past, present and future. Hum Hered. 2008;65:209–220. [DOI] [PubMed] [Google Scholar]
- 37.Jorde LB, Shortsleeve PA, Henry JW, VanBuren RT, Hutchinson LE, Rigley TM. Genetic analysis of the Utah population: a comparison of STR and VNTR loci. Hum Biol. 2000;72:927–936. [PubMed] [Google Scholar]
- 38.VanderWeele TJ, Ding P. Sensitivity analysis in observational research: Introducing the “E-Value”. Ann Intern Med. 2017;167(4):268–274. [DOI] [PubMed] [Google Scholar]
- 39.America’s Families and Living Arrangements: 2014. Current Population Survey, 2014 Annual Social and Economic Supplement. https://www.census.gov/data/tables/2014/demo/families/cps-2014.html. Retrieved on April 18, 2020
- 40.Hassold T, Chiu D. Maternal age-specific rates of numerical chromosome abnormalities with special reference to trisomy. Hum Genet. 1985;70:11–17. [DOI] [PubMed] [Google Scholar]
- 41.Reddy UM, Wapner RJ, Rebar RW, Tasca RJ. Infertility, assisted reproductive technology, and adverse pregnancy outcomes: executive summary of a National Institute of Child Health and Human Development workshop. Obstet Gynecol. 2007;109:967–977. [DOI] [PubMed] [Google Scholar]
- 42.Natoli JL, Ackerman DL, McDermott S, Edwards JG. Prenatal diagnosis of Down syndrome: a systematic review of termination rates (1995–2011). Prenat Diagn. 2012;32:142–153. [DOI] [PubMed] [Google Scholar]
- 43.Sommers BD, Long SK, Baicker K. Changes in mortality after Massachusetts health care reform. A quasi-experimental study changes in mortality after Massachusetts health care reform. Ann Intern Med. 2014;160:585–593. [DOI] [PubMed] [Google Scholar]
- 44.Parker SE, Mai CT, Canfield MA, et al. Updated national birth prevalence estimates for selected birth defects in the United States, 2004–2006. Birth Defects Res A Clin Mol Teratol. 2010;88:1008–1016. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.