OBJECTIVE: To examine the influence of fetal and maternal characteristics on the survival of children born with Down syndrome.
METHODS: We used prospectively collected population-based data on Down syndrome extracted from the UK Northern Congenital Abnormality Survey (NorCAS), January 1, 1985–December 31, 2003, matched to hospital and national mortality records to January 29, 2008, to determine survival status of liveborn children. Survival to 20 years was estimated by using Kaplan-Meier methods. Cox proportional hazards regression was used to examine factors that predict survival.
RESULTS: A total of 1115 Down syndrome pregnancies were notified to NorCAS during the 19 years, a total prevalence of 16.8 (95% CI, 15.8–17.8) per 10 000 live births and stillbirths. Of these, 5.4% resulted in a spontaneous fetal loss (late miscarriage ≥20 weeks and stillbirth), 31.7% in a termination of pregnancy, and 63.0% in a live birth. Survival status was known for 95.3% of live births; 16.6% resulted in a death. Year of birth (P < .001), gestational age at delivery (P < .001), standardized birth weight (P < .001), karyotype (P < .01), and presence of additional structural anomalies (P < .001) were significant predictors of survival. Infant gender, plurality, maternal age, and maternal deprivation were not significant predictors of survival.
CONCLUSIONS: These robust estimates of predictors of survival are important for the prenatal counseling of parents whose pregnancy is affected by Down syndrome and for health care planning for the future care needs of these children.
- aHR —
- adjusted hazard ratio
- CI —
- confidence interval
- HR —
- hazard ratio
- IMD —
- Index of Multiple Deprivation
- NHS —
- National Health Service
- NorCAS —
- Northern Congenital Abnormality Survey
- ONS —
- Office for National Statistics
- PMS —
- Northern Perinatal Mortality Survey
- VLBW —
- very low birth weight
What’s Known on This Subject:
Survival of children born with Down syndrome has been improving, but few studies have used population-based data to examine the influence of fetal and maternal characteristics on survival.
What This Study Adds:
This study examined predictors of survival for children born with Down syndrome using population-based data from the UK Northern Congenital Abnormality Survey and shows that year of birth, gestational age, birth weight, and presence of additional anomalies influence survival status.
Down syndrome is a chromosomal congenital anomaly caused by an additional copy of all, or part, of chromosome 21. It is the most common nonlethal chromosomal anomaly, with a reported prevalence, in the United States, of 1.3 per 1000 live births.1 The live birth prevalence of Down syndrome in a specific population is determined by both the maternal age distribution and the availability and uptake of prenatal diagnosis and termination of pregnancy. For those parents who decide to continue pregnancies affected by Down syndrome, advances in fetal and neonatal care have improved outcomes for children born with Down syndrome such that the majority of children can now be expected to reach adulthood.2–8 However, there are limited data on the influence of a range of fetal and maternal characteristics that may be important predictors of survival, including the presence of additional structural anomalies. Such information is needed to inform parents and health professionals of prognoses and to guide commissioning for both the future health care needs and the educational and social care requirements of affected individuals and their families.
This study used prospectively collected, population-based register data, matched with 3 high-quality sources of mortality information, to investigate the influence of a range of fetal and maternal characteristics on the survival status of individuals born with Down syndrome over an 19-year period.
The North of England, UK (Fig 1), is a geographically distinct area with a stable population of 3 million and ∼32 000 deliveries per year. Data on all structural and chromosomal congenital anomalies occurring within this population are collected by the Northern Congenital Abnormality Survey (NorCAS) and include those arising in late miscarriages (fetal deaths at 20–23 weeks’ gestation), terminations of pregnancy for fetal anomaly after prenatal diagnosis at any gestation, stillbirths (fetal deaths delivered at ≥24 weeks’ gestation), and live births to mothers resident in the region, regardless of their delivery location. Cases are ascertained from multiple sources, including antenatal ultrasound, fetal medicine departments, cytogenetic laboratories, the regional cardiology center, pathology departments, and pediatric surgery, and are included when first diagnosed up to age 16 years (1985–2001) or, more recently, age 12 years (2001–present). Further details of data collection have been previously published.9 The NorCAS is a member of the British Isles Network of Congenital Anomaly Registers10 and the European Surveillance of Congenital Anomalies.11
All anomalies are coded according to the World Health Organization International Statistical Classification of Diseases and Related Health Problems 10th Revision12 and categorized into congenital anomaly group (the system affected), subtype (the specific condition), and syndrome (where applicable) according to the European Surveillance of Congenital Anomalies guidelines.13 Up to 6 congenital anomalies can be recorded for each individual.
All children with Down syndrome (including trisomy 21, Q90.0, mosaic Down syndrome, Q90.1, and translocations involving chromosome 21 Q90.2) who were delivered between January 1, 1985, to December 31, 2003, formed this population-based case series.
Information on Deaths
The Northern Perinatal Mortality Survey (PMS) is a population-based survey of all perinatal and infant deaths in the North of England.14 The NorCAS and PMS are held in a common database at the Regional Maternity Survey Office in Newcastle and linked via the mother’s details. The PMS holds details of all deaths in the first year of life. Information on deaths occurring in subsequent years, and for those who moved out of the region before the age of 1 year, was obtained from the UK Office for National Statistics (ONS). Individuals were traced by National Health Service (NHS) number (where available), maternal surname, infant gender, date of birth, and last known address. Deaths were matched individuals with a corresponding mortality record, whereas survivors were matched individuals with no record of death before January 28, 2008 (the date that the trace was carried out). Those individuals who were not found on the PMS or matched with records from the ONS, were identified from information held by NorCAS or hospital records. Survival status for those identified was established using the NHS National Strategic Tracing Service. Cases that could not be traced by any of these methods were excluded from the analysis.
Prevalence rates were estimated as the total number of cases per 10 000 births (live births and stillbirths). The 95% confidence intervals (CIs) for prevalence rates were derived from the binomial distribution. Temporal trends in prevalence (or proportion) were examined by the Cochran-Armitage test. Temporal trends for other variables were examined by the Cuzick test for trend.
Surviving children with Down syndrome were censored with their age on January 28, 2008. Survival in the general UK population was estimated from published life tables.15 The cumulative probability of a spontaneous fetal loss was also estimated by the Kaplan-Meier method. Gestational age at delivery was used as the time scale variable and ‘failure’ was defined as any spontaneous fetal death, including miscarriages (20–23 weeks gestation age) and stillbirths (≥24 weeks gestation). Cases delivered as elective terminations of pregnancy were censored at their gestational age of delivery, while cases delivered as live births were treated as having survived until the latest gestational age. Retaining live births in the denominator provides a more accurate estimate of the true probability of spontaneous fetal loss by avoiding distortions from late stillbirths at very rare gestational ages.
The following factors were examined as potential influences of survival: karyotype (full trisomy 21, mosaicism, and translocation), year of birth (1985–1990, 1991–1996, and 1997–2003), infant gender (male and female), plurality (singleton and multiple), gestational age at delivery (<30, 30–36, and ≥37 weeks), maternal age at delivery (<20, 20–30, and >30 years), birth weight, index of multiple deprivation (IMD), and presence of additional structural anomalies. Birth weight was standardized for gestational age, gender, and parity (z score <1, between −1 and 1, and >1) by applying a customized fetal growth formula to regional birth weight references.16 The IMD 2004, an area-based measure of socioeconomic deprivation,17 was determined for each case using the mother's postal code at delivery and was analyzed as within-region tertiles. The presence of additional anomalies was classified into 5 categories: no additional anomalies, cardiovascular anomalies only, digestive system anomalies only, cardiovascular and digestive system anomalies only, or any other combination of anomalies.
Survival within each variable category, and 95% CIs were derived by the Kaplan-Meier method, with differences between categories being examined by using the log-rank test. To prevent presentation of unreliable estimates derived from the tail of the Kaplan-Meier curve, survival probabilities were not determined where, or once, there were <10 at-risk cases. In these instances, the probability of survival to a given age was approximated by using Cox proportional hazards regression. Cox proportional hazards regression was also used to examine the independent influence of each variable by constructing an adjusted model containing all variables on test with year of birth, gestational age, and standardized birth weight included as continuous variables. Interaction terms were used to examine whether survival had improved similarly among cases with and without additional anomalies. The Grambsch-Therneau test for constancy of the log hazard ratio (HR) function was used to examine adherence to the proportional hazards assumption.
Statistical analyses were carried out by using Stata 10.1 (StataCorp, College Station, TX). P < .05 was considered statistically significant.
NorCAS, as part of the British Isles Network of Congenital Anomaly Registers network, is exempt from requiring consent for inclusion on the register under section 251 of the NHS Act 2006 and has ethics approval (09/H0405/08) to undertake studies involving the data. A favorable ethical opinion for this study was given by the South Tees Local Research Ethics Committee.
A total of 1115 pregnancies affected by Down syndrome were notified to the NorCAS during the 19 years, giving a total prevalence of 16.8 (95% CI, 15.8–17.8) per 10 000 live and stillbirths. Total prevalence increased from 14.0 (95% CI, 12.5–15.6) per 10 000 births in 1985–1990 to 20.4 (95% CI, 18.5–22.4) per 10 000 births in 1997–2003 (P test for trend <.001). There were 1058 (94.9%) pregnancies with trisomy 21, 34 (3.1%) with a translocation affecting chromosome 21, and 23 (2.1%) with mosaic Down syndrome (Table 1). There were 24 (2.2%) twin or higher order pregnancies affected by Down syndrome; 21 (87.5%) of these resulted in a live birth (Table 1).
There were 645 (57.9%) male and 469 (42.1%) female total Down syndrome pregnancies, giving a male-to-female ratio of 1:0.73 (Table 1). Similar proportions of male and female pregnancies resulted in a fetal loss, termination of pregnancy, or live birth (P = .41). There were 738 (66.2%) pregnancies in women aged >30 years. Median maternal age increased from 30 years (95% CI, 29–31) in 1985–1990 to 34 years (95% CI, 34–35) in 1997–2003 (P test for trend <.001); the proportion of Down syndrome pregnancies to older mothers (≥30 years) increased from 52.5% (95% CI, 47.0%–57.9%) in 1985–1990 to 77.2% (95% CI, 73.2%–81.1%) in 1997–2003 (P test for trend <.001).
Four hundred eighty-nine (43.9%) Down syndrome pregnancies occurred in women from the region's least deprived tertile (Table 1). Also, 535 (48%) pregnancies were delivered at term (≥37 weeks) and 110 (9.9%) babies had a birth weight z score >1.
In 416 (37.3%) pregnancies, there was one or more additional structural anomaly (Table 2). In 348 (31.2%) Down syndrome pregnancies, there was a coincident cardiovascular anomaly; 308 (88.5%) had only a cardiovascular anomaly, 25 (7.2%) had a cardiovascular and digestive anomaly. and 15 (4.3%) had a cardiovascular and other anomaly. The most common cardiovascular subtypes were atrioventricular septal defects (n = 155, 13.9%) and ventricular septal defects (n = 112, 10%). There were digestive system anomalies in 57 (5.1%) pregnancies; 28 (49.1%) pregnancies had only a digestive system anomaly, 25 (43.9%) pregnancies had a digestive system and cardiac anomaly, and 4 (7.0%) had a digestive system anomaly and another structural anomaly (but not cardiac). Duodenal atresia occurred in 36 (32.2%) pregnancies with Down syndrome. Other additional structural anomalies included anomalies of the urinary system (n = 11, 1.1%) and nervous system (n = 7, 0.6%) (Table 2).
Sixty (5.4%) pregnancies resulted in a spontaneous fetal death and 353 (31.7%) in a termination of pregnancy for fetal anomaly (Table 1). The overall probability of a spontaneous fetal loss, accounting for case censoring, was 7.9% (95% CI, 6.2%–10.0%). The proportion of pregnancies ending in termination of pregnancy increased from 18.0% (95% CI, 13.9%–22.6%) in 1985–1990 to 39.0% (95% CI, 34.4%–43.8%) in 1997–2003 (P < .001). There were 702 (63.0%) live births, giving a live birth prevalence of 10.6 (95% CI, 9.9%–11.5%) per 10 000 live births. Live birth prevalence was 11.1 (95% CI, 9.8–12.5) per 10 000 live births in 1985–1990, 9.4 (95% CI, 8.1–10.8) per 10 000 live births for 1991–1996, and 11.5 (95% CI, 10.1–13.0) per 10 000 live births in 1997–2003 (P for trend = .81).
Survival status was known for 669 (95.3%) Down syndrome live births. Of the traced cases, 111 (16.6%) resulted in a death; survival status was unknown for 33 (4.7%) individuals with Down syndrome. Seventy-eight deaths occurred in the first year of life, giving an infant mortality rate of 11.7% (95% CI, 9.3%–14.3%). Infant mortality decreased significantly from 14.0% (95% CI, 9.9%–19.2%) in 1985–1990 to 5.8% (3.2%–9.6%) in 1997–2003 (P = .002). Table 3 and Fig 2 show survival for children live born with Down syndrome over the study period; 10- and 20- year survival rates were 83.9% and 82.9%, respectively. Ten-year survival increased from 78.3% in 1985–1990 to 91.2% in 1997–2003. There were few deaths from 10 to 20 years; 20-year survival increased from 77.5% in 1985–1990 to a predicted (by Cox proportional hazards regression) 90.7% in 1997–2003.
There were no deaths among individuals with trisomy 21 mosaicism; survival was significantly better compared with individuals with full trisomy 21 (HR indeterminate, medium unbiased rate ratio = 0.21 [95% CI, 0.00–0.93]; P = .04). No apparent difference was observed in the survival of children with Down syndrome resulting from full trisomy and children with Down syndrome resulting from a translocation (Table 3) (adjusted HR [aHR] = 0.97 [95% CI, 0.30–3.10]; P = .60).
Year of birth was a significant predictor of survival (aHR, per year = 0.89 [95% CI, 0.85–0.92]; P < .001; Fig 1) with survival highest in the later epoch (1997–2003) (Table 3). There was a significant influence of gestational age at delivery (aHR = 0.76 [95% CI, 0.72–0.80]; P < .001) on survival; only 23.1% (95% CI, 5.6%–47.5%) of children with Down syndrome born at <30 weeks’ gestation survived to 1 year compared with 92.0% (95% CI, 89.2%–94.1%) of those born at term. Increasing standardized birth weight also positively predicted survival (aHR, per z score = 0.80 [95% CI, 0.71–0.91]; P < .001). There was a nonsignificant relationship between being large for gestational age and survival (aHR = 0.64 [95% CI, 0.27–1.53]; P = .32).
The presence of additional anomalies increased mortality (Table 3, Fig 2); with a 5- to 8-fold increased risk of mortality in the presence of a cardiovascular, digestive, or both cardiovascular and digestive system anomalies. The survival of children with both cardiovascular and a digestive system anomalies was not significantly lower than for those with digestive system anomalies only (P = .27). There was no evidence that temporal improvements in survival were different between cases with and without additional anomalies (Fig 2). The 10-year survival for children born with Down syndrome in 1997–2003 with no additional structural anomalies was 99.1% (95% CI, 93.4%–99.9%), not significantly less than the general population of the United Kingdom during 2000 (P = .78).
There was little evidence that infant gender, plurality, maternal age, or IMD were significant predictors of survival (P > .2) (Table 3).
This large, population-based study estimated the influence of a range of fetal and maternal characteristics on survival in children born with Down syndrome over a 19-year study period by using data from an established high-quality congenital anomaly registry. There were no deaths among individuals with mosaic trisomy 21, and survival was similar between individuals with trisomy 21 arising from a translocation and those with trisomy 21. Year of birth, gestational age at delivery, and standardized birth weight were independent positive predictors of survival. The presence of additional anomalies significantly reduced survival; however, survival was improved over time for all individuals with Down syndrome, regardless of the presence of additional anomalies. Infant gender, plurality, maternal age, and maternal deprivation were not found to significantly influence survival.
Our study has several strengths. We have used data on Down syndrome derived from a longstanding, high-quality congenital anomaly registry. The NorCAS ascertains cases from multiple sources including cytogenetic laboratories and undertakes regular validation with the Northern Genetics Service. The NorCAS contributes to established UK and European networks that use similar inclusion criteria and have a consistent approach to data collection, coding, and recording. NorCAS includes notifications of cases up to 12 years of age. Death information was obtained from death registrations held by the UK ONS. This information is very complete18 and will include all deaths including those that occur before surgery has taken place. This is likely to explain our lower survival rates for children with Down syndrome and a cardiac anomaly compared with a recent study from the same region that examined survival using information based solely on postsurgical mortality.19 Although primarily a study of postnatal survival, we have also included pregnancies ending in spontaneous fetal loss and termination of pregnancy for fetal anomaly after prenatal diagnosis to give a complete picture of the natural history of Down syndrome–affected pregnancies.
The study also has some limitations. Despite the use of 3 comprehensive sources of information on deaths, we were unable to trace almost 5% of individuals with Down syndrome. It is likely that a proportion of these children will have been adopted, making tracing difficult. For some variables, we did not have enough liveborn cases in each category to observe the survival, so less accurate approximations (using Cox regression modeling), were used. We used an area-based measure of deprivation derived from the maternal residential postal code rather than individual measures such as maternal educational attainment or parental occupation, which were not available to the study. We have accurate information on the date of death, but information on the cause of death was not available to the study. However, previous studies have shown that the main causes of death in children born with Down syndrome are cardiac in origin, infection, and leukemia.20
The total prevalence of Down syndrome in the region during this 19-year period was 16.8 per 10 000 births. Total prevalence has increased during the study period; 2 contributors to this increase are advancing maternal age and increased prenatal detection of cases that would have been undiagnosed fetal losses. Live birth prevalence remained constant, reflecting a balance between increasing prenatal detection and elective termination of pregnancy.
We found a male preponderance among total Down syndrome pregnancies but no significant difference in the proportion of live births for females compared with the proportion of live births for males. Infant gender did not influence survival. Some previous studies have also reported no difference in survival by gender.6,20,21 However, Glasson et al’s study from Western Australia reported a longer life expectancy among males with Down syndrome (61.1 years) compared with females (57.8 years), which is contrary to that in the general population.22
There was a significant association between year of birth and the probability of survival, with survival increasing over time. In the United Kingdom, diagnostic testing for Down syndrome has been an integral part of routine antenatal screening for over three decades. Universal screening has been offered since April 2004.23 The increase in survival over time is likely to result from a number of improvements in medical care including surfactant therapy and corticosteroids for respiratory distress syndrome,24 intrapartum treatment of chorioamnionitis to prevent neonatal sepsis,25 public health campaigns that have lowered the incidence of sudden unexpected death in infancy,26 and advances in echocardiography and cardiac surgery.27 Overall survival for individuals with Down syndrome is still below that of the general UK population. However, our results suggest the majority of this is due to the presence of additional anomalies; indeed, we found the 10-year survival rate for children born with Down syndrome and no additional structural anomalies was similar to that of the general population.
Maternal age is a well-known risk factor for Down syndrome, and we found that two-thirds of Down syndrome pregnancies occurred in mothers >30 years old. Nearly half of these mothers chose to have a termination of pregnancy, reflecting the age-related screening policies available in the United Kingdom during the study period as opposed to universal screening policies now in place. However, maternal age at delivery was not a significant predictor of survival, as has been found by others.7
Gestational age at delivery and standardized birth weight were also found to be significant predictors of survival. In 2010, Boghossian et al28 found that very low birth weight (VLBW) children born with Down syndrome were at an increased risk of death compared with VLBW children without Down syndrome. Without survival information for VLBW infants without Down syndrome in the general population, we were not able to confirm this.
As other studies have reported,29,30 we found a higher prevalence of Down syndrome pregnancies in women living in the least deprived parts of the region. More women from the least deprived areas had a termination of pregnancy compared with women from the most deprived areas. Nevertheless, we found no association between deprivation and survival; although this may be due to limited study power or that the IMD, despite being a standard estimate of deprivation used in the United Kingdom, did not adequately assess the relevant individual-level exposures.
Survival for children born with Down syndrome has increased over time. However, there is little available population-based data on the influence of a range of fetal and maternal characteristics on survival status. We have found that year of birth, gestational age at delivery, standardized birth weight, and presence of additional structural anomalies predict survival status. It is essential that parents are presented with accurate information on expected outcome during prenatal counseling and that health, education, and social care planning for the future needs of these children and their families is based on high-quality data. However, while improving survival is essential, further research on subjective quality of life is also needed if we are to attempt to address the many challenges facing individuals living with Down syndrome.
We thank all the Link clinicians in the North of England for their continued support of the NorCAS. NorCAS is funded by the UK Healthcare Quality Improvement Partnership.
- Accepted January 25, 2012.
- Address correspondence to Judith Rankin, PhD, Institute of Health & Society, Baddiley-Clarke Building, Newcastle University, Richardson Rd, Newcastle upon Tyne, NE2 4AX, UK. E-mail:
Prof Rankin conceived the study, designed the study with Dr Pearce and wrote the paper; Ms Bythell and Prof Rankin contributed to acquisition of the data; Mr Tennant undertook the data analysis and contributed to drafting the paper; and all authors contributed to the interpretation of the data and gave final approval for the paper to be published.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: This study was funded by the UK charity the Newlife Foundation for Disabled Children.
- Dastgiri S,
- Gilmour WH,
- Stone DH
- Hayes C,
- Johnson Z,
- Thornton L,
- et al
- ↵British Isles Network of Congenital Anomaly Registers. Available at: www.binocar.org/. Accessed March 19, 2012
- ↵European Surveillance of Congenital Anomalies. Available at: www.eurocat-network.eu/. Accessed on March 19, 2012
- ↵World Health Organization. International statistical classification of diseases and related health problems 10th revision: version for 2007. Available at: www.who.int/classifications/icd10/. Accessed March 19, 2012
- ↵European Surveillance of Congenital Anomalies. Coding of EUROCAT subgroups of congenital anomalies. Available at: www.eurocat-network.eu/content/EUROCAT-Definition-New-Subgroups-Feb-2007.pdf. Accessed March 19, 2012
- Hey EN,
- Welch RG,
- Lawson JB,
- et al
- ↵World Health Organization. Life tables for WHO member states. Available at: www.who.int/healthinfo/statistics/mortality_life_tables/en/. Accessed March 19, 2012
- Noble M,
- Wright G,
- Dibben C,
- et al
- ↵Irving CA, Chaudhari MP. Cardiovascular abnormalities in Down’s syndrome: spectrum, management and survival over 22 years [published online ahead of print August 11, 2011]. Arch Dis Child. doi:10.1136/adc.2010.210534
- UK National Screening Committee
- Perloff JK,
- Warnes CA
- Boghossian NS,
- Hansen NI,
- Bell EF,
- et al.,
- Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network
- Vrijheid M,
- Dolk H,
- Stone D,
- Abramsky L,
- Alberman E,
- Scott JE
- Copyright © 2012 by the American Academy of Pediatrics