Increased parity and risk of trisomy 21: review of 37 110 live births

BMJ 1997; 314 doi: (Published 08 March 1997) Cite this as: BMJ 1997;314:720
  1. MS Schimmel, senior neonatologista,
  2. AI Eidelman, professor of pediatricsa,
  3. C Hammerman, senior neonatologista,
  4. E Kornbluth, social workera,
  5. P Zadka, senior epidemiologistb
  1. a Department of Neonatology Shaare Zedek Medical Center PO Box 3235 Jerusalem 91031 Israel
  2. b Health Division, Israel Central Bureau of Statistics, Jerusalem, Israel
  1. Correspondence to: Professor Eidelman
  • Accepted 15 November 1996


Routine amniocentesis for all women over the age of 35 for detecting trisomy 21 has been the world standard for nearly two decades. It is based on a calculated tradeoff: the documented increased risk of trisomy 21 with advancing age versus the risk of fetal loss secondary to the amniocentesis procedure.

Recent data suggest that more specific individual estimates of risk can be made with a combination of maternal age and serum markers.1 We have reported a trend towards an increased risk of trisomy 21 with increased parity.2 Given the uniqueness of our population—ultra-Orthodox Jewish mothers who have a high rate of delivery over the age of 35, high parity, and who abstain from prenatal screening and abortion—we had an opportunity to study the interrelation of age and parity.

Patients, methods, and results

We reviewed all deliveries at the Shaare Zedek Medical Center in 1981-9, the period before Embedded Image fetoprotein and ultrasound screening for the prenatal detection of trisomy 21 became available in Israel. All clinically suspected cases of Down's syndrome were cytogenetically confirmed as trisomy 21. Mothers were analysed by five year age groups and further subdivided by parity as primiparous, multiparous (second to fifth delivery), and grand multiparous (sixth delivery or more). Statistical comparison between groups was by Poisson distribution analysis. The attributable risk of parity was calculated by a standard formula.3

In 37 110 live births, trisomy 21 was diagnosed in 54 infants, resulting in an overall incidence of 1.46 per 1000 live births (1:687 live births). Table 1) shows the rate of trisomy 21 in each five year age group and the prevalence of trisomy 21 in each age group by parity. Grand multiparity was associated with an increased prevalence in the age groups between 25 and 44 (P<0.005 for ages 25-29, P<0.08 for ages 30-34, P<0.001 for ages 40-44), and prevalence was higher in multiparous women than in primiparous women in all age groups. In the 40-44 age group, prevalence was significantly higher with grand multiparity than multiparity (P<0.001). In the 20-24 age group, prevalence of trisomy 21 was significantly higher for multiparous mothers than for primiparous women and for the average rate of this group (P<0.05). The attributable risk of parity to the birth of an infant with trisomy 21 was calculated to be 15%.

Table 1

Prevalence of trisomy 21 at Shaare Zedek Medical Center, Israel, 1981-9. Values are rate (per 1000 live births) and ratio, unless specified otherwise

View this table:


In the past five years the reliance on maternal age alone as an indication for amniocentesis for detecting fetal trisomy 21 has been modified by the use of multiple maternal serum markers.4 Our data show that parity can be used to help identify mothers under 35 at high risk and mothers over 35 at low risk.

Because the relatively small size of our sample (37 110 deliveries and 54 infants with trisomy 21) precluded analysis in each yearly age cohort group and for each parity level, the analysis was performed for 5 year age groups and by conventional categories of parity. In almost all age groups with enough grand multiparous mothers for a valid calculation, these groups had a significantly higher rate of trisomy 21 than did all the mothers in the group. A ratio of 1:38 in the grand multiparity 40-44 age group and the absence of any infants with trisomy in 4299 primiparous women aged 25-40 strongly suggests an interrelation between age and parity, calculated to be 15% greater than from age related factors alone.

Other, much smaller studies do not necessarily support our data.5 Further population studies are needed to provide a large database for exact calculations that will provide a basis for public health recommendations and guidelines for individual families.


Funding: No additional funding.

Conflict of interest: None.


Presented in part at the annual meeting of the Society for Pediatric Research, Seattle, May 1994.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract