DOI: 10.1111/1471-0528.

14269 Maternal medicine

www.bjog.org

Pregnancy at very advanced maternal age: a UK

population-based cohort study
KE Fitzpatrick,a D Tuffnell,b JJ Kurinczuk,a M Knighta
a
National Perinatal Epidemiology Unit, Nuffield Department of Population Health, University of Oxford, Oxford, UK b Department of
Obstetrics and Gynaecology, Bradford Teaching Hospitals NHS Foundation Trust, Bradford, UK
Correspondence: M Knight, National Perinatal Epidemiology Unit, University of Oxford, Old Road Campus, Oxford OX3 7LF, UK. Email
[email protected]

Accepted 18 July 2016. Published Online 1 September 2016.

Objectives To describe the characteristics, management and women appeared more likely than comparison women to have
outcomes of women giving birth at advanced maternal age pregnancy complications including gestational hypertensive
(≥48 years). disorders, gestational diabetes, postpartum haemorrhage, caesarean
delivery, iatrogenic and spontaneous preterm delivery on
Design Population-based cohort study using the UK Obstetric
univariable analysis and after adjustment for demographic and
Surveillance System (UKOSS).
medical factors. However, adjustment for multiple pregnancy or use
Setting All UK hospitals with obstetrician-led maternity units. of assisted conception attenuated most effects, with significant
associations remaining only with gestational diabetes (adjusted odds
Population Women delivering at advanced maternal age
ratio [aOR] 4.81, 95% CI 1.93–12.00), caesarean delivery (aOR 2.78,
(≥48 years) in the UK between July 2013 and June 2014 (n = 233)
95% CI 1.44–5.37) and admission to an intensive care unit (aOR
and 454 comparison women.
33.53, 95% CI 2.73–412.24).
Methods Cohort and comparison group identification through the
Conclusions Women giving birth at advanced maternal age have
UKOSS monthly mailing.
higher risks of a range of pregnancy complications. Many of the
Main outcome measures Pregnancy complications. increased risks appear to be explained by multiple pregnancy or
use of assisted conception.
Results Older women were more likely than comparison women to
be overweight (33% versus 23%, P = 0.0011) or obese (23% versus Keywords Advanced maternal age, assisted reproduction, cohort
19%, P = 0.0318), nulliparous (53% versus 44%, P = 0.0299), have study, pregnancy, pregnancy outcomes.
pre-existing medical conditions (44% versus 28%, P < 0.0001), a
Tweetable abstract The pregnancy complications in women giving
multiple pregnancy (18% versus 2%, P < 0.0001), and conceived
birth aged 48 or over are mostly explained by multiple pregnancy.
following assisted conception (78% versus 4%, P < 0.0001). Older

Please cite this paper as: Fitzpatrick KE, Tuffnell D, Kurinczuk JJ, Knight M. Pregnancy at very advanced maternal age: a UK population-based cohort study.
BJOG 2016;124:1097–1106.

Many studies have reported an association between

Introduction
advanced maternal age and a higher risk of adverse mater-
Childbearing at advanced maternal age is becoming nal and infant outcomes.4–6 However, the majority of stud-
increasingly common in high-income countries.1,2 Further- ies have reported outcomes in women aged ≥35 years or
more, developments in artificial reproductive technologies, women aged ≥40 years. These studies therefore include
such as ovum donation, may contribute to an increasing only a small number of the oldest mothers and have not
incidence of pregnancies in women outside the usual bio- specifically assessed the risks in women of very advanced
logical reproductive age. In England and Wales the average maternal age, in whom adverse outcomes could be more
age at childbearing has increased steadily since the mid- common. The small numbers of studies that have specifi-
1970s from 26.4 in 1975 to 30.0 in 2013, with a corre- cally investigated outcomes in relation to very advanced
sponding rise in the proportion of women delivering in maternal age7 have largely not made any attempt to control
their thirties and forties.3 for potential confounding factors and have predominately

ª 2016 The Authors. BJOG An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd on behalf of 1097
Royal College of Obstetricians and Gynaecologists.
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use,
distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.
 Fitzpatrick et al.

been conducted using retrospective review of medical sociodemographic factors; model 2 additionally adjusted for
records over a number of years in a single or small number previous medical history; and model 3 additionally adjusted
of institutions. Such studies suffer from a number of limi- for relevant pregnancy-related factors. ‘Missing’ was
tations such as limited generalisability and lack of statistical included as an extra category for variables that had ≥10%
power. The objective of this national population-based missing data. Continuous variables were tested for evidence
study was to describe the characteristics, management and of departure from linearity by the addition of first-order
outcomes of women giving birth at very advanced maternal fractional polynomials to the model and subsequent likeli-
age in the UK and to estimate the risk of adverse outcomes hood ratio testing. Continuous variables that showed evi-
attributable to very advanced maternal age. dence of nonlinearity were treated and presented as
categorical in the analysis, whereas those showing evidence
of linearity were treated as continuous linear terms when
Methods
adjusting for them in the analysis but presented as categori-
A national, population-based cohort study was conducted. cal for ease of interpretation. Plausible interactions were
The cohort included any pregnant woman in the UK at tested in the full regression model by the addition of interac-
20 weeks of gestation or more, who was of very advanced tion terms and subsequent likelihood ratio testing on
maternal age. Although very advanced maternal age has gen- removal, with a P-value <0.01 considered as evidence of sig-
erally been used to refer to women aged ≥45 years, for prag- nificant interaction to account for multiple testing.
matic reasons, and so as to not over-burden reporting Women who initially had a multiple pregnancy but then
clinicians, we defined very advanced maternal age as women had fetal reduction were classified in the analysis according
aged ≥48 years at their date of delivery. The cohort was iden- to the number of fetuses left after the reduction. Sponta-
tified through the monthly mailing of the UK Obstetric neous first-trimester losses in women known initially to
Surveillance System (UKOSS) between 1 July 2013 and 30 have a multiple pregnancy were classified in the analysis
June 2014. The UKOSS methodology has been described in according to the post-loss number of fetuses. Second-trime-
detail elsewhere.8 Briefly, cards were sent to nominated clini- ster losses in a multiple pregnancy were classified according
cians (midwives, risk management midwives, obstetricians to the pre-loss number of fetuses in the main analysis, but
and anaesthetists) in each of the UK’s obstetrician-led mater- were not included when examining neonatal outcomes
nity units requesting the number of pregnant woman of very unless they occurred after 24 weeks. Logistic regression
advanced maternal age they had seen that month. On report- using robust standard errors to allow for non-independence
ing a pregnancy in a woman of very advanced maternal age, of neonates from multiple births was used when comparing
clinicians were sent a data collection form to complete seeking neonatal outcomes.
additional information concerning the characteristics, man- Using the most recent national birth data3,9,10 we antici-
agement and outcomes of the woman concerned. Reporting pated identifying 406 women aged ≥48 years at their date
clinicians were also asked to identify and complete an identi- of delivery and 812 comparison women. With these num-
cal data collection form for comparison women, defined as bers of women the study would have had an estimated
the two pregnant women at 20 weeks of gestation or more power of 80% at the 5% level of statistical significance to
who were <48 years of age at their estimated date of delivery detect odds ratios of ≥1.5 and ≥2.0, assuming outcomes
and who delivered immediately before the older woman in have an incidence of 40% and 5%, respectively. The actual
the same hospital. All data requested were anonymous. Infor- number of older and comparison women identified during
mation on woman’s year of birth and expected date of deliv- the study gave an estimated power of 80% at the 5% level
ery was used to identify duplicate reports. A total of five of significance to detect odds ratios of ≥1.6 and ≥2.5,
reminders were sent if complete forms were not returned. assuming the same outcome incidence levels.
All analyses were conducted using STATA v 13 software
(Statacorp, College Station, TX, USA). Odds ratios (ORs)
Results
with 95% CIs were estimated throughout using uncondi-
tional logistic regression. Odds ratios were adjusted for All eligible hospitals with obstetrician-led maternity units
potential confounding factors if there was a pre-existing contributed data to UKOSS during the study period (100%
hypothesis or evidence that the factors were potential con- response), notifying 351 women of very advanced maternal
founders or mediators of the relationship between age. Excluding those subsequently reported by clinicians as
advanced maternal age and the outcome in question. To not fulfilling the inclusion criteria, data collection forms
help examine the relative influence of the potential con- were obtained for 89% of the notified women (Figure S1)
founders and mediators on the association between mater- and data were received for 454 comparison women. A total
nal age and the outcome in question, models were adjusted of 233 women of very advanced maternal age were identi-
in a hierarchical fashion: model 1 adjusted for fied.

1098 ª 2016 The Authors. BJOG An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd on behalf of
Royal College of Obstetricians and Gynaecologists
 Pregnancy at advanced maternal age

The median age of the older women was 49 years (range and Wales on women giving birth at very advanced mater-
48–61 years) whereas the median age of the comparison nal age (Table 2) with the exception of the proportion of
women was 31 years (range 16–46 years) (Figure 1). Older women who were married/in a civil partnership with one
women were significantly more likely than comparison or more previous live-born children, which was lower in
women to be overweight or obese, to be nonsmokers, to our study population.
have had previous uterine surgery not including previous Older women were significantly more likely than the
caesarean section, to have previous or pre-existing medical comparison women to have a plan at booking for more
condition(s), to be nulliparous, to have a multiple preg- than the recommended number of antenatal visits for low-
nancy, and to have conceived following assisted conception risk women 82% (185/226) versus 30% (133/447),
(Table 1). Of the 50 older women who conceived without P < 0.001). Maternal age was the commonest reason given
assisted conception, 14 (61% of the 23 in whom this was for older women having more antenatal visits (82%, 152/
known) had planned pregnancies. Of the 176 older women 185), followed by underlying medical condition/previous
known to have conceived following assisted conception, obstetric history (32%, 60/185). The proportion of older
51% (61/119, 57 women with no information provided) women who received care at their usual hospital for their
had the assisted conception performed outside the UK, place of residence was not significantly different from the
91% (137/151, 25 with no information provided) had used proportion seen in the comparison women (86%, 197/228,
egg donation, 21% (22/104, 72 with no information pro- versus 91%, 409/451, P = 0.089). Of the older women who
vided) had used sperm donation, and 97% (147/152, 24 did not have their care at their usual hospital, reasons
with no information provided) underwent in vitro fertilisa- included patient preference (n = 20), referral to a tertiary
tion/intracytoplasmic injection (IVF/ICSI). Of the 147 centre because of underlying medical conditions (n = 4)
women who had IVF/ICSI, 55 women had the number of and maternal age (n = 2). Eighty-seven percent of older
embryos transferred recorded: 22 had one embryo trans- women (196/225) had antenatal screening; 74% had a
ferred, 25 had two embryos transferred, six had three nuchal translucency test, 57% had serum screening, 1%
embryos transferred and two had four embryos transferred. had chorionic villus sampling, 5% had amniocentesis and
Excluding first-trimester spontaneous losses and subsequent 77% had an 18- to 20-week anomaly scan. The proportions
fetal reductions, 15 of the 30 women who had more than seen in the comparison women were comparable with the
one embryo transferred went on to have a multiple preg- exception of amniocentesis, which was significantly less
nancy. The characteristics of the older women in our study likely in the comparison women, occurring in just 1%
were comparable to the available national data for England (P = 0.009). Older women were also more likely than the

Figure 1. Characteristics of older and comparison women.

ª 2016 The Authors. BJOG An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd on behalf of 1099
Royal College of Obstetricians and Gynaecologists
 Fitzpatrick et al.

Table 1. Characteristics of older and comparison women

Characteristic Number (%)* of Number (%)* of Unadjusted OR P-value

older women (n = 233) comparison (95% CI)
women (n = 454)

Sociodemographic characteristics
Ethnic group
White 165 (71) 323 (71) 1
Non-White 67 (29) 129 (29) 1.02 (0.72–1.44) 0.9259
Marital status
Married or cohabiting 194 (85) 375 (84) 1
Single 35 (15) 71 (16) 0.95 (0.61–1.48) 0.8299
Socio-economic group
Managerial and professional occupations 91 (39) 144 (32) 1
Other 103 (44) 231 (51) 0.71 (0.50–1.00) 0.0511
Missing 39 (17) 79 (17)
Body mass index (kg/m2)
<25 101 (44) 260 (58) 1
25–29.9 75 (33) 103 (23) 1.87 (1.29–2.73) 0.0011
≥30 52 (23) 85 (19) 1.57 (1.04–2.38) 0.0318
Smoking status
Never/ex smoker 226 (99) 407 (90) 1
Smoked during pregnancy 3 (1) 45 (10) 0.12 (0.04–0.39) 0.0004
Previous medical history
Previous uterine surgery not including previous caesarean section
No 168 (74) 418 (93) 1
Yes 60 (26) 33 (7) 4.52 (2.85–7.17) <0.0001
Previous or pre-existing medical condition
No 129 (56) 328 (72) 1
Yes 101 (44) 126 (28) 2.04 (1.46–2.84) <0.0001
Pregnancy-related characteristics
Parity
0 122 (53) 200 (44) 1
1 or more 108 (47) 252 (56) 0.7 (0.51–0.97) 0.0299
Previous caesarean section
No 179 (79) 379 (84) 1
Yes 49 (21) 72 (16) 1.44 (0.96–2.16) 0.0765
Multiple pregnancy
No 189 (82) 444 (98) 1
Yes 41 (18) 10 (2) 9.63 (4.73–19.63) <0.0001
Conceived following assisted conception
No 50 (22) 425 (96) 1
Yes 176 (78) 19 (4) 78.74 (45.13–137.38) <0.0001

*Percentage of individuals with complete data unless missing category shown.

comparison women to have a third-trimester ultrasound comparison women to have a range of complications
performed (87%, 199/228 versus 50%, 222/445, P < 0.001). including gestational hypertensive disorders, gestational
Reasons for a third-trimester ultrasound included concerns diabetes, postpartum haemorrhage, caesarean delivery,
about the fetus, e.g. fetal growth (n = 113), routine diabetic iatrogenic and spontaneous preterm delivery and intensive
monitoring (n = 15), other maternal condition (n = 9), therapy unit (ITU) admission. With the exception of gesta-
abnormal presentation (n = 5) and pregnancy complication tional diabetes, caesarean delivery and ITU admission,
(n = 17). however, these effects were attenuated and became non-
Table 3 shows the pregnancy complications experienced significant largely after adjustment for pregnancy-related
by the older and comparison women. Unadjusted analysis characteristics. For effects that became nonsignificant after
suggests that older women were more likely than adjustment for pregnancy-related characteristics, further

1100 ª 2016 The Authors. BJOG An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd on behalf of
Royal College of Obstetricians and Gynaecologists
 Pregnancy at advanced maternal age

pregnancies but two spontaneously lost one twin in the

Table 2. Comparison of the characteristics of older women
identified by UKOSS with available national data on women giving first trimester, two spontaneously lost one twin in the sec-
birth at very advanced maternal age ond trimester before 24 weeks of gestation, and one lost
both twins in the second trimester. Three other older
Characteristic UKOSS National P-value
women experienced spontaneous loss in the second trime-
number (%)* data** on
of older number (%)*
ster before 24 weeks: one woman initially had a triplet
women of older pregnancy but spontaneously lost one of the triplets in the
(n = 233) women second trimester before 24 weeks, and the other two were
(n = 384) singleton pregnancies. Among the older women, this effec-
tively left a total of 268 fetuses surviving beyond 24 weeks
Marital status of gestation (35 sets of twins, three sets of triplets and 189
Married/civil partnership 127 (55) 246 (64) singletons). Of these 268 fetuses, three were stillborn
Single 102 (29) 138 (36) 0.0754 antepartum: one was a set of twins where both twins were
Number of previous live-born children within marriage/civil
stillborn. A further two of the fetuses died shortly after
partnership
birth following very preterm delivery (<28 weeks of gesta-
0 67 (53) 94 (38)
1 or more 60 (47) 152 (62) 0.0056 tion), equating to an overall perinatal mortality rate of 18.7
Multiple pregnancy per 1000 (95% CI 6.1–42.9). This was more than three
No 189 (82) 332 (86) times the national rate of 5.5 per 100011 (relative risk 3.33,
Yes 41 (18) 52 (14) 0.185 95% CI 1.40–7.93). The perinatal mortality rate among sin-
Age of mother (years) gletons was 15.9 per 1000 (95% CI 3.3–45.7), also statisti-
48 73 (31) 139 (36) 0.2041
cally significantly higher than the national perinatal
49 55 (24) 100 (26) 0.5793
mortality rate among singletons of 5.2 per 100011 (relative
≥50 105 (45) 145 (38) 0.0861
risk 3.03, 95% CI 0.99–9.33, P = 0.043).
*Percentage of individuals with complete data. The proportion of fetuses surviving beyond 24 weeks of
**Data for maternities in England and Wales 2013, Office of gestation that had a congenital anomaly was similar
National Statistics ad hoc data and analysis 2015.
between the older women and the comparison women
(1.9%, 5/263 versus 1.5%, 7/460, P = 0.702), as was the
analysis was performed in which adjustment was made for proportion that had other major complications such as res-
sociodemographic factors, previous medical history and rel- piratory distress syndrome and severe infection (2%, 4/205
evant pregnancy-related factors one at a time (Table S1); versus 3.8%, 13/344, P = 0.240). The proportion of fetuses
while parity and previous caesarean section, where relevant, that had a low birthweight (<2500 g) was higher among
had little impact, all of the effects became nonsignificant those born to older women compared with comparison
after adjustment for how the woman conceived and some women (32%, 85/267 versus 8%, 38/463, P < 0.001),
became nonsignificant following adjustment for multiple although this difference disappeared after controlling for
pregnancy. An analysis including just singleton pregnancies gestational age at delivery.
(Model 4, Table 3) reflected very similar results with the
exception of postpartum haemorrhage, which remained sig-
Discussion
nificant even after full adjustment. There was evidence of
significant interaction between caesarean delivery and par- Main findings
ity: the raised odds of having a caesarean delivery were only This study suggests that women giving birth at very
apparent in nulliparous older women (adjusted OR [aOR] advanced maternal age have a higher risk of having a range
9.90, 95% CI 3.64–26.92 in nulliparous women; aOR 0.71, of pregnancy complications including gestational hyperten-
95% CI 0.31–1.66 in parous women). No other significant sive disorders, gestational diabetes, postpartum haemor-
interactions were found. Among the older women who had rhage, caesarean delivery, iatrogenic and spontaneous
a caesarean delivery, maternal age was the primary indica- preterm delivery and ITU admission. With the exception of
tion for 21% (36/175). Other indications included fetal gestational diabetes, caesarean delivery and ITU admission,
compromise (19%, 33/175), maternal compromise (14%, these increased risks appear to be largely explained by the
25/175), failure to progress (14%, 24/175), abnormal pre- higher rate of multiple pregnancy or use of assisted concep-
sentation (10%, 18/175), previous caesarean section (9%, tion observed in the older women.
16/175) and maternal request (5%, 9/175).
Two of the older women had fetal reduction, one from Strengths and limitations
three to two fetuses and the other from three to one fetus. A key strength of our study is its prospective population-
Five of the older women were known to initially have twin based national design, which reduces the possibility of bias

ª 2016 The Authors. BJOG An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd on behalf of 1101
Royal College of Obstetricians and Gynaecologists
 Table 3. Pregnancy complications of older and comparison women in all women and in women with singleton pregnancies only

1102
Number (%)* Number (%)* Unadjusted P-value Model 1 Model 2 Model 3 Model 4
of older of comparison OR (95% CI)
women women Adjusted P-value Adjusted P-value Adjusted P-value Adjusted P-value
(n = 233) (n = 454) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
Fitzpatrick et al.

Any gestational hypertensive disorder

No 196 (85) 430 (95) 1 1 1 1 1
Yes 34 (15) 24 (5) 3.11 (1.79–5.38) <0.0001 2.88 (1.63–5.09) 0.0003 2.84 (1.60–5.06) 0.0004 2.13 (0.75–6.02) 0.1535 1.82 (0.56–5.99) 0.3221
Any gestational hypertensive disorder managed by early delivery
No 219 (95) 442 (97) 1 1 1 1 1
Yes 11 (5) 12 (3) 1.85 (0.80–4.26) 0.1482 1.72 (0.74–4.01) 0.2078 1.8 (0.77–4.22) 0.1783 1.28 (0.27–6.01) 0.7536 1.49 (0.27–8.28) 0.6462
Pregnancy induced hypertension
No 209 (91) 440 (97) 1 1 1 1 1
Yes 21 (9) 14 (3) 3.16 (1.57–6.33) 0.0012 2.85 (1.38–5.88) 0.0046 2.81 (1.35–5.84) 0.0058 2.82 (0.79–10.05) 0.1093 2.53 (0.63–10.21) 0.1912
Preeclampsia
No 217 (94) 444 (98) 1 1 1 1 1
Yes 13 (6) 10 (2) 2.66 (1.15–6.16) 0.0225 2.55 (1.07–6.07) 0.0346 2.53 (1.05–6.09) 0.0379 1.16 (0.22–6.09) 0.8577 0.82 (0.10–6.44) 0.8483
Gestational diabetes
No 188 (82) 436 (96) 1 1 1 1 1
Yes 42 (18) 18 (4) 5.41 (3.04–9.65) <0.0001 4.97 (2.73–9.04) <0.0001 4.78 (2.61–8.77) <0.0001 4.81 (1.93–12.00) 0.0007 3.41 (1.27–9.19) 0.0151
Gestational diabetes requiring insulin
No 221 (96) 452 (100) 1 1 1 1 1
Yes 9 (4) 2 (0) 9.2 (1.97–42.96) 0.0047 8.12 (1.72–38.31) 0.0081 7.54 (1.57–36.17) 0.0115 3.64 (0.50–26.55) 0.2033 3.85 (0.49–30.18) 0.1999
Placenta praevia
No 222 (97) 453 (100)
Yes 8 (3) 0 (0)
Placental abruption
No 226 (99) 451 (100) 1 1 1 1 1
Yes 3 (1) 2 (0) 2.99 (0.50–18.04) 0.2316 4.7 (0.65–34.09) 0.1258 4.31 (0.60–30.89)¥ 0.1456 1.2 (0.08–19.02)# 0.8963 1.11 (0.07–18.70)¥,# 0.941
Diagnosed postpartum haemorrhage
No 169 (74) 385 (85) 1 1 1 1 1
Yes 59 (26) 69 (15) 1.95 (1.32–2.88) 0.0009 1.89 (1.26–2.84) 0.0022 1.74 (1.13–2.66)¥ 0.0114 2.03 (0.97–4.27)# 0.0603 2.45 (1.10–5.44)¥,# 0.0279
Diagnosed postpartum haemorrhage requring blood transfusion
No 210 (94) 441 (98) 1 1 1 1 1
Yes 14 (6) 8 (2) 3.67 (1.52–8.90) 0.0039 3.53 (1.39–8.92) 0.0077 2.30 (0.87–6.05)¥ 0.0916 4.33 (0.94–19.96)# 0.06 6.39 (1.19–34.42)¥,# 0.0308
Thrombotic event
No 229 (100) 453 (100)
Yes 0 (0) 1 (0)
Labour induced
No 156 (69) 321 (71) 1 1 1 1 1
Yes 71 (31) 133 (29) 1.1 (0.78–1.55) 0.5945 1.18 (0.82–1.69) 0.3711 1.1 (0.75–1.61)¥ 0.6235 1.91 (1.03–3.54)# 0.0401 1.9 (0.98–3.65)¥,# 0.056
Caesarean delivery
No 50 (22) 305 (67) 1 1 1 1 1
Yes 178 (78) 149 (33) 7.29 (5.03–10.55) <0.0001 6.41 (4.39–9.37) <0.0001 5.9 (3.98–8.75)¥ <0.0001 2.78 (1.44–5.37)# 0.0024 2.98 (1.47–6.05)¥,# 0.0024

ª 2016 The Authors. BJOG An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd on behalf of
Royal College of Obstetricians and Gynaecologists
 Table 3. (Continued)

Number (%)* Number (%)* Unadjusted P-value Model 1 Model 2 Model 3 Model 4
of older of comparison OR (95% CI)
women women Adjusted P-value Adjusted P-value Adjusted P-value Adjusted P-value
(n = 233) (n = 454) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)

Gestational age at delivery (weeks)

Term (37+ 176 (78) 420 (93) 1 1 1 1 1
weeks)
Iatrogenic 32 (14) 17 (4) 4.49 (2.43–8.30) <0.0001 4.49 (2.39–8.43) <0.0001 4.23 (2.19–8.18)¥ <0.0001 1.01 (0.30–3.45)# 0.9845 1.72 (0.40–7.34) ¥,# 0.4671
preterm

Royal College of Obstetricians and Gynaecologists

(<37 weeks)
Spontaneous 18 (8) 17 (4) 2.53 (1.27–5.02) 0.0081 2.44 (1.17–5.09) 0.0169 2.34 (1.09–5.00)¥ 0.0287 1.11 (0.28–4.45)# 0.8832 0.75 (0.14–4.15) ¥,# 0.7431
preterm (<37
weeks)
Admitted to ITU
No 224 (97) 453 (100) 1 1 1 1 1
Yes 6 (3) 1 (0) 12.13 (1.45–101.40) 0.0212 10.98 (1.28–94.01) 0.0288 10.96 (1.28–94.17) 0.0291 33.53 (2.73–412.24) 0.0061 37.53 (2.98–472.18) 0.005
In nulliparous
women
Caesarean delivery
No 1
Yes 9.90 (3.64–26.92) <0.001
In women
parity 1+
Caesarean delivery
No 1
Yes 0.71 (0.31–1.66) 0.431

Model 1: Adjusted for socio-demographic factors (Ethnic group, marital status, socio-economic group, body mass index and smoking status).
Model 2: Adjusted for variables included in model 1 plus previous medical history (previous uterine surgery not including previous caesarean section and previous or preexisting medical
conditions where ¥ is shown, or just previous or preexisting medical conditions where ¥ is not shown).
Model 3: Adjusted for variables included in model 2 plus pregnancy related factors (parity, multiple pregnancy, how conceived and previous caesarean delivery where # is shown, or just
parity, multiple pregnancy and how conceived if # is not shown.
Model 4: Women with singleton pregnacies only, adjusted for socio-demographic factors (Ethnic group, marital status, socio-economic group, body mass index and smoking status) plus
previous medical history (previous uterine surgery not including previous caesarean section and previous or preexisting medical conditions where ¥ is shown, or just previous or preexisting
medical conditions where ¥ is not shown) plus pregnancy related factors (parity, how conceived and previous caesarean delivery where # is shown, or just parity and how conceived if # not

ª 2016 The Authors. BJOG An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd on behalf of
shown).
Statistically significant values are bolded.
*Percentage of individuals with complete data.
Pregnancy at advanced maternal age

1103
 Fitzpatrick et al.

associated with hospital-based studies. We also had the ≥45 years,7,13 but generally higher than rates quoted in con-
advantage of not relying on coded data from routine hospital temporary studies for women of more modest advanced
administrative systems, which has been shown to have a num- maternal age (≥35 or ≥40 years).15–20 Nearly one in five
ber of limitations.12 Despite the active monthly nature of the women giving birth at very advanced maternal age in our
UKOSS data collection system and the presence of several study developed gestational diabetes, around five-fold higher
reporting clinicians in each hospital, comparison with the than the rate in the comparison group with differences per-
most recent national birth data, which does not cover the sisting after adjustment for potential confounding and medi-
entire study period, suggests that we have under-ascertained ating factors. This rate is in the higher range of the rates that
women giving birth at very advanced maternal age by up to have been reported.7,13,15–20 Our rate of postpartum haemor-
30%. Comparison of the characteristics of women giving rhage, diagnosed in just over one-quarter of women of very
birth at very advanced maternal age in our study with the advanced maternal age, was also higher than those
available national data for England and Wales on women giv- quoted.13,16–18,20,21 These differences may reflect disparities in
ing birth at very advanced maternal age suggested that our estimating and defining postpartum haemorrhage.22
study may have under-ascertained parous older women, A recent systematic review identified a higher risk of cae-
although other characteristics are comparable. However, sarean section among women of advanced maternal age
adjustment of our results for parity did not have a meaningful (mainly ≥35 years), although the heterogeneity among the
impact, suggesting that this is unlikely to have had a substan- included studies precluded a pooled estimate of the risk.23
tive effect. Our caesarean section rate is at the higher end of the range
of those reported for women aged ≥45 years,7,13 and may
Interpretation (in light of other evidence) reflect a tendency for clinicians to offer caesarean delivery in
Our finding that women giving birth aged 48 years and this extreme age group. Indeed, advanced maternal age was
older have a higher risk of having a range of pregnancy the commonest indication, recorded as the primary reason
complications is comparable to the limited number of for around one-fifth of the caesarean deliveries.
studies that have assessed outcomes in high-income Just over one in five women of very advanced mater-
countries in relation to very advanced maternal age nal age in our study delivered preterm, with the rates of
(≥45 years).7,13 Of particular note is the fact that the both iatrogenic and spontaneous preterm delivery higher
odds of the majority of these complications were attenu- than in the comparison women, differences that appear
ated or disappeared after adjustment for mode of con- to be largely explained by differences in pregnancy-
ception and multiple pregnancy, which has important related characteristics. Most of the previous literature has
implications for counselling and practice in assisted reported an association between advanced maternal age
reproduction services. Half of the women who had dou- and preterm delivery without separating out type of pre-
ble embryo transfer went on to have a multiple preg- term delivery; our estimated rate of preterm delivery is
nancy, which is higher than the rate of 29% reported in in the higher range of the rates reported for women
a meta-analysis of trials of double embryo transfer,14 aged ≥45 years7,13 and is generally higher than in studies
potentially due to the use of ovum donation in these examining women of more modestly advanced maternal
women. It cannot therefore be assumed that multiple age.15,18,19,24,25 Except for the higher rate of low birth-
pregnancy is less likely in this population than in women weight infants, which appears to be largely linked to the
of younger ages undergoing assisted reproduction. Rec- high preterm delivery rate, other infant outcomes were
ommendations regarding assisted conception including comparable between the older and comparison women in
egg donation in older mothers, as well as single embryo our study. However, the perinatal mortality rate was sig-
transfer should take these findings into account. nificantly raised in the older women in comparison to
Adjustment for medical co-morbidities had little impact the national rate. Other studies have reported an increase
on the odds ratios we observed, suggesting that the in the risk of perinatal mortality among women of
increased risk of pregnancy complications is unlikely to be advanced maternal age, although the absolute increase in
solely due to the population of women who undergo risk appears to be small.7,13,25,26
assisted reproduction being less healthy than those who
conceive spontaneously. Nearly one in six women giving
Conclusion
birth at very advanced maternal age in our study developed
a gestational hypertensive disorder with nearly 1 in 10 devel- Although having a baby at very advanced maternal age is
oping pregnancy-induced hypertension and just over 1 in 20 currently uncommon in the UK, developments in artificial
developing pre-eclampsia. Although these rates were around reproductive technologies are contributing to an increasing
three-fold higher than in the comparison women, they are in incidence of pregnancies in women outside of the normal
the lower range of the rates reported for women aged reproductive age. Women giving birth at very advanced

1104 ª 2016 The Authors. BJOG An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd on behalf of
Royal College of Obstetricians and Gynaecologists
 Pregnancy at advanced maternal age

maternal age have a higher risk of having a range of preg- Acknowledgements

nancy complications in comparison to younger women, The authors would like to thank the UK Obstetric Surveil-
including a higher risk of gestational hypertensive disorders, lance System (UKOSS) reporting clinicians who notified
gestational diabetes, postpartum haemorrhage, caesarean cases and completed the data-collection forms, without
delivery, iatrogenic and spontaneous preterm delivery and whose contribution it would not have been possible to
ITU admission. With the exception of gestational diabetes, carry out this study.
caesarean delivery and ITU admission, these increased risks
appear to be largely explained by the higher rate of multiple
pregnancy or the use of assisted conception observed in the Supporting Information
older women, all of which are inextricably inter-related to Additional Supporting Information may be found in the
older maternal age, with older age leading to a need for IVF online version of this article:
if conception is to occur and age itself and IVF leading to an Figure S1. Case reporting and completeness of data col-
increased risk of multiple birth. These findings should be lection.
considered when counselling and managing women of very Table S1. Further analysis of pregnancy complications,
advanced maternal age. They also show the implications for adjusting for sociodemographic factors, previous medical
maternity services of having a baby at very advanced mater- history and relevant pregnancy-related factors one at a
nal age. There may be a place for considering fetal reduction time. &
in women of very advanced age with multiple pregnancies
although the long-term effect of fetal reduction on surviving
infants is unclear and this needs further research. Recom- References
mendations regarding assisted conception including egg 1 Mathews TJ, Hamilton BE. Mean age of mother, 1970–2000. Natl
donation in older mothers, as well as single embryo transfer, Vital Stat Rep 2002;51:1–13.
should take these findings into account. 2 Breart G. Delayed childbearing. Eur J Obstet Gynecol Reprod Biol
1997;75:71–3.
3 Office for National Statistics. Live Births In England and Wales by
Disclosure of interests Characteristics of Mother 1, 2013. London: Office for National
None declared. Completed disclosure of interests form Statistics, 2014.
available to view online as supporting information. 4 Balasch J, Gratacos E. Delayed childbearing: effects on fertility and the
outcome of pregnancy. Curr Opin Obstet Gynecol 2012;24:187–93.
5 Montan S. Increased risk in the elderly parturient. Curr Opin Obstet
Contribution to authorship
Gynecol 2007;19:110–2.
KEF designed the study, coded the data, carried out the 6 Hansen JP. Older maternal age and pregnancy outcome: a review of
analysis and wrote the manuscript. DT and JJK assisted the literature. Obstet Gynecol Surv 1986;41:726–42.
with the design of the study and contributed to the writing 7 Carolan M. Maternal age ≥45 years and maternal and perinatal
of the manuscript. MK conceived and designed the study outcomes: a review of the evidence. Midwifery 2013;29:479–89.
and contributed to the writing of the manuscript. MK will 8 Knight M, Kurinczuk JJ, Tuffnell D, Brocklehurst P. The UK obstetric
surveillance system for rare disorders of pregnancy. BJOG
act as guarantor. 2005;112:263–5.
9 General Register Office for Scotland. Vital Events Reference Tables
Details of ethical approval 2013. Edinburgh: General Register Office for Scotland, 2014.
The North London Research Ethics Committee 1 approved 10 Northern Ireland Statistics and Research Agency. Registrar General
Annual Report 2013. Belfast: Northern Ireland Statistics and
the study (reference 10/H0717/20) on 06/04/2010.
Research Agency, 2014.
11 Manktelow BM, Smith LK, Evans TA, Hyman-Taylor P, Kurinczuk JJ,
Funding Field DJ, et al. Perinatal Mortality Surveillance Report UK Perinatal
This article presents independent research funded by the Deaths for births from January to December 2013. Leicester: The
National Institute for Health Research (NIHR) under the Infant Mortality and Morbidity Group, Department of Health
‘Beyond maternal death: improving the quality of mater- Sciences, University of Leicester, 2015.
12 Knight M. Studies using routine data versus specific data collection:
nity care through national studies of “near-miss” mater- what can we learn about the epidemiology of eclampsia and the
nal morbidity’ programme (programme grant RP-PG- impact of changes in management of gestational hypertensive
0608-10038). Marian Knight is funded by an NIHR disorders? Pregnancy Hypertens 2011;1:109–16.
Research Professorship. The views expressed in this pub- 13 Carolan MC, Davey MA, Biro M, Kealy M. Very advanced maternal
age and morbidity in Victoria, Australia: a population based study.
lication are those of the author(s), and not necessarily
BMC Pregnancy Childbirth 2013;13:80.
those of the NHS, the NHIR or the Department of 14 McLernon DJ, Harrild K, Bergh C, Davies MJ, de Neubourg D,
Health. The funder had no role in the study design, data Dumoulin JC, et al. Clinical effectiveness of elective single versus
collection and analysis, decision to publish, or prepara- double embryo transfer: meta-analysis of individual patient data
tion of the article. from randomised trials. BMJ 2010;341:c6945.

ª 2016 The Authors. BJOG An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd on behalf of 1105
Royal College of Obstetricians and Gynaecologists
 Fitzpatrick et al.

15 Ciancimino L, Lagana AS, Chiofalo B, Granese R, Grasso R, Triolo O. 21 Callaway LK, Lust K, McIntyre HD. Pregnancy outcomes in women
Would it be too late? A retrospective case–control analysis to of very advanced maternal age. Aust N Z J Obstet Gynaecol
evaluate maternal–fetal outcomes in advanced maternal age. Arch 2005;45:12–6.
Gynecol Obstet 2014;290:1109–14. 22 Rath WH. Postpartum hemorrhage – update on problems of
16 Khalil A, Syngelaki A, Maiz N, Zinevich Y, Nicolaides KH. Maternal definitions and diagnosis. Acta Obstet Gynecol Scand
age and adverse pregnancy outcome: a cohort study. Ultrasound 2011;90:421–8.
Obstet Gynecol 2013;42:634–43. 23 Bayrampour H, Heaman M. Advanced maternal age and the risk of
17 Biro MA, Davey MA, Carolan M, Kealy M. Advanced maternal age cesarean birth: a systematic review. Birth 2010;37:219–26.
and obstetric morbidity for women giving birth in Victoria, Australia: 24 Kenny LC, Lavender T, McNamee R, O’Neill SM, Mills T, Khashan
a population-based study. Aust N Z J Obstet Gynaecol AS. Advanced maternal age and adverse pregnancy outcome:
2012;52:229–34. evidence from a large contemporary cohort. PLoS One 2013;8:
18 Yogev Y, Melamed N, Bardin R, Tenenbaum-Gavish K, Ben-Shitrit G, e56583.
Ben-Haroush A. Pregnancy outcome at extremely advanced 25 Carolan M, Frankowska D. Advanced maternal age and adverse perinatal
maternal age. Am J Obstet Gynecol 2010;203:558 e1–7. outcome: a review of the evidence. Midwifery 2011;27:793–801.
19 Cleary-Goldman J, Malone FD, Vidaver J, Ball RH, Nyberg DA, 26 Huang L, Sauve R, Birkett N, Fergusson D, van Walraven C.
Comstock CH, et al. Impact of maternal age on obstetric outcome. Maternal age and risk of stillbirth: a systematic review. CMAJ
Obstet Gynecol 2005;105:983–90. 2008;178:165–72.
20 Carolan M, Davey MA, Biro MA, Kealy M. Older maternal age and
intervention in labor: a population-based study comparing older and
younger first-time mothers in Victoria, Australia. Birth 2011;38:24–9.

1106 ª 2016 The Authors. BJOG An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd on behalf of
Royal College of Obstetricians and Gynaecologists