Archives of Gynecology and Obstetrics

, Volume 285, Issue 4, pp 951–957

Perinatal correlates of delayed childbearing in a developing country

Authors

    • Department of Community Health and Primary Care, College of MedicineUniversity of Lagos
    • Institute of Child Health and Great Ormond Street Hospital for Children NHS TrustUniversity College London
  • O. A. Solanke
    • Consultant Obstetrician and GynecologistLagos Island Maternity Hospital
Maternal-Fetal Medicine

DOI: 10.1007/s00404-011-2105-5

Cite this article as:
Olusanya, B.O. & Solanke, O.A. Arch Gynecol Obstet (2012) 285: 951. doi:10.1007/s00404-011-2105-5

Abstract

Objective

To evaluate women with delayed childbearing after 35 years and the perinatal profile of their surviving offspring in a low-income country.

Method

A matched case–control study of advanced maternal age (AMA) parturients above 35 years was compared to younger mothers (20–35 years) in an inner-city maternity hospital in southwest Nigeria using conditional logistic regression analysis.

Results

A total of 513 were enlisted as cases matched with 1,539 controls. AMA was associated with marital status, occupation, parity, antenatal care and elective cesarean delivery but significantly less likely to be associated with human-immunodeficiency-virus infection and cephalopelvic disproportion. After adjusting for these maternal factors, AMA was not associated with any adverse perinatal outcomes.

Conclusion

Delayed childbearing after 35 years was generally not associated with adverse pregnancy outcomes in this setting and may be indicative of an appreciable awareness of the potential pregnancy risks resulting in higher uptake of obstetric services by AMA mothers.

Keywords

Delayed childbearingPregnancy outcomesElective cesarean deliveryAntenatal careSub-Saharan Africa

Introduction

Delayed childbearing at advanced maternal age (AMA >35 years) is underpinned by a constellation of voluntary and involuntary factors that include but not limited to late (re)marriage, delayed conception from infertility, academic and career pursuits, desire for large family, lack of or ineffective family planning and longer life expectancy [1, 2]. The contributions of these factors to delayed childbearing often differ between women in developed and developing countries. For example, academic and career pursuits appear to be more common in developed countries; thus the affected women are more likely to be primiparous [3]. In contrast, childbearing at AMA is more common among multiparous women in developing countries as a result of factors such as lack of or ineffective family planning underpinned by favorable cultural disposition towards large family size and poverty [4].

Several studies have reported adverse pregnancy outcomes such as fetal growth restriction, preterm birth, low birth weight and perinatal mortality due to delayed childbearing [1, 2, 5, 6], thus warranting special treatment for AMA parturients as a “high-risk” group [7]. In contrast, other studies found no such evidence especially when pre-existing medical disorders or prior reproductive problems are either absent or effectively addressed [3, 810]. A systematic review of studies in developed countries conducted between 1985 and 2002 concluded that there was insufficient evidence to determine if older maternal age was an independent risk factor for preterm and small-for-gestational age birth [11]. These conflicting findings are not confined to developed countries as demonstrated by similar studies from developing countries [1218]. However, there appears to be some consensus that AMA is associated with elevated risks for chromosomal abnormalities, malpresentation, spontaneous abortion, preterm labor and cesarean delivery [2, 14, 19, 20].

Against the backdrop of the varied reports on perinatal outcomes among surviving offspring of AMA mothers especially in developing countries this study therefore set out to determine the perinatal correlates of AMA in a Sub-Saharan African community after adjusting for potentially confounding maternal factors.

Methods

Study design and population

This retrospective case–control study was conducted at the Island Maternity Hospital (IMH), a state-owned tertiary hospital that serves as a referral centre for several private and public hospitals in an inner-city community in metropolitan Lagos, southwest Nigeria. Participants were drawn from a previously reported cohort recruited under a UNHS program [21, 22]. In brief, all surviving newborns delivered between May 2005 and December 2007 were recruited and screened for sensorineural hearing loss. Newborns whose mothers were too ill to be interviewed including near-misses and those who died during childbirth as well as those who died shortly (<24 h) after enrolment were excluded. All mother–infant pairs who participated in the primary study were considered eligible for this current study. From this group, all consecutive AMA mothers (>35 years) were recruited as cases along with younger mothers aged 20–35 years (designated “YMA”) as controls. Cases and controls were matched by date of delivery and infant’s gender at the ratio of 1:3. Although post-hoc power analysis for retrospective studies is a subject of debate, an indication of whether the study sample was sufficiently powered to provide reliable estimates of statistically significant differences between cases and controls was attempted to complement the evidence from the confidence intervals of the estimates [23]. Thus, assuming approximately 6.0% exposure risk among AMA mothers and 2.5% exposure risk among YMA mothers with an odds ratio of 2.50, a sample size of 510 cases and 1,530 controls will have 90% power to detect a 3.5% difference in exposure risk between the two groups at 95% confidence interval. If the effect of case–control matching was considered, the required sample size would be lower than 2,040. Ethical approval was obtained from Lagos State Health Management Board, Nigeria and University College London, UK [21]. Informed consent was obtained from all participants prior to enrolment.

Study variables

Selected factors were guided by evidence from the literature [1, 2, 6, 10, 1218]. Non-clinical information was obtained predominantly from the mothers at enrolment while clinical data was obtained from available hospital records. Included were maternal socio-demographic variables such as ethnicity, religion, education and occupation as well as pregnancy-related factors such as parity, antenatal care (at least one visit or none), human immunodeficiency virus (HIV) status, previous cesarean section, hypertensive disorders (inclusive of pre-eclampsia, eclampsia and pregnancy-induced hypertension), antepartum hemorrhage, placenta previa, premature rupture of membranes, cephalopelvic disproportion, malpresentation (cephalic or non-cephalic), prolonged/obstructed labor, and mode of delivery (vaginal, elective cesarean or emergency cesarean delivery). Parity was classified as “primiparity”, “multiparity” (parity 2–4) and grand-multiparity (parity >4). Perinatal outcomes consisted of gestational type (singletons or multiple), gestational age (based on maternal recall of last menstrual period), birth weight, fetal growth restriction (based on birthweight below 2 SD of the mean birthweight for each gestational age from a previously validated fetal growth curve for this study population [24]), fetal distress, suspected birth asphyxia (indexed by low Apgar scores at 1 and 5 min), possible sepsis (used collectively for septicemia, meningitis and pneumonia), hyperbilirubinemia (requiring phototherapy) and admission into special care baby unit (SCBU) and mode of feeding (exclusive breast feeding, bottle only or mixed).

Statistical analysis

After a descriptive overview of the factors among study participants matched and unmatched odds ratios (OR) and the corresponding 95% confidence intervals (CI) were determined with Pearson χ2 test or Fisher’s exact test as appropriate. Perinatal outcomes associated with AMA were identified through separate conditional and unconditional multivariable logistic regression models adjusting for significant maternal confounders. Socio-demographic and obstetric factors with p ≤ 0.10 from univariate analyses were considered as maternal confounders and covariates for the pregnancy outcomes in the logistic regression analysis. Potential interaction of pairs of variables entered into the logistic models was assessed with likelihood ratio test and model calibration was verified with the Hosmer–Lemeshow test. All tests were two-sided at a critical level of p < 0.05 and SPSS for Windows version 16.0 (SPSS Inc, Chicago, IL, USA) was used for all statistical analyses.

Results

Of the 4,575 participating mother–infant pairs over the study period, a total of 513 (11.2%) mothers were older than 35 years and recruited as cases while 1,539 mothers (20–35 years) were enlisted as controls. The median age of cases was 38 (inter-quartile range 37–39) years as compared to 29 (inter-quartile range 26–31) years for the controls. The controls were consecutively chosen within a range of 5 days of delivery date of each case after excluding all teenage mothers (76 or 1.7%) and mother–infant pairs (324 or 7.1%) with incomplete data.

The socio-demographic and obstetric characteristics of the participants are presented in Table 1. More than half of the mothers recruited for this study belonged to the Yoruba ethnic tribe, were Christians, had formal education and either self-employed or engaged in formal employment. A quarter of AMA mothers were primiparous and 13.1% were grand-multiparous. Some 28.8% of AMA mothers also did not receive antenatal care while 15.4% delivered by elective cesarean section. All but three of AMA mothers were married. Slightly over half (52.4%) of the offspring were male. In the unmatched and matched univariate analyses, AMA was associated with married status (OR: 3.44), informal/self-employment (OR: 2.91) or formal/full-time employment (OR: 2.99), multiparity (OR: 2.97), grand-multiparity (OR: 10.83), antenatal care (OR: 1.43) and delivery by elective cesarean section (OR: 1.58). Older mothers were also significantly less likely to have cephalopelvic disproportion (OR: 0.50). Considering that parity emerged with the highest odds, a stratified analysis of maternal factors among AMA women showed that multiparous mothers were significantly likely to be Muslims, have non-tertiary education, self-employed or engaged in full-time/formal employment and delivered vaginally but less likely to have cephalopelvic disproportion (data not shown). They were also more likely to exclusively breastfeed their babies. Multiparous AMA mothers were significantly more likely to attend antenatal care than multiparous YMA mothers (p = 0.004). However, there was no statistically significant difference between primiparous and grand-multiparous AMA as compared to their younger counterparts.
Table 1

Socio-demographic and obstetric factors associated with advanced maternal age

Factors

Cases (%), n = 513

Controls (%), n = 1539

Unadjusted odds ratio (95% CI)

Unmatched

Matched

Marital status

 Unmarried

3 (0.6)

31 (2.0)

1.0

1.0

 Married

510 (99.4)

1,508 (98)

3.50 (1.06–11.48)*

3.44 (1.05–11.27)*

Ethnicity

 Yoruba

399 (77.8)

1164 (75.6)

1.0

1.0

 Hausa

14 (2.7)

38 (2.5)

1.08 (0.58–2.00)

1.08 (0.58–2.01)

 Ibo and others

100 (19.5)

337 (21.9)

0.86 (0.67–1.11)

0.87 (0.67–1.11)

Religion

 Christian

299 (58.3)

918 (59.6)

1.0

1.0

 Others

214 (41.7)

621 (40.4)

1.06 (0.86–1.30)

1.06 (0.86–1.30)

Education

 None

12 (2.3)

29 (1.9)

1.0

1.0

 Primary/Secondary

313 (61.0)

902 (58.6)

0.84 (0.42–1.66)

0.84 (0.42–1.68)

 Tertiary

188 (36.6)

608 (39.5)

0.75 (0.37–1.49)

0.75 (0.37–1.50)

Occupation

 None

39 (7.6)

301 (19.6)

1.0

1.0

 Informal/self-employment

291 (56.7)

770 (50.0)

2.92 (2.04–4.18)***

2.91 (2.02–4.18)***

 Full-time/formal employment

183 (31.6)

468 (30.4)

3.02 (2.08–4.39)***

2.99 (2.05–4.37)***

Parity

 Primiparous

130 (25.3)

830 (53.9)

1.0

1.0

 2–4

316 (61.6)

671(43.6)

3.01 (2.39–3.78)***

2.97 (2.35–3.74)***

 >4

67 (13.1)

38 (2.5)

11.26 (7.26–17.46)***

10.83 (6.90–17.00)***

Antenatal care

 None

148 (28.8)

561 (36.5)

1.0

1.0

 Yes

365 (71.2)

978 (63.5)

1.42 (1.14–1.76)**

1.43 (1.14–1.78)**

Maternal HIV-positive

 No

495 (96.5)

1,441 (93.6)

1.0

1.0

 Yes

18 (3.5)

98 (6.4)

0.54 (0.32–0.89)*

0.54 (0.32–0.90)*

Previous cesarean section

 No

448 (87.3)

1,371 (89.1)

1.0

1.0

 Yes

65 (12.7)

168 (10.9)

1.18 (0.87–1.61)

1.19 (0.87–1.62)

Hypertensive disorders

 No

473 (92.2)

1,438 (93.4)

1.0

1.0

 Yes

40 (7.8)

101 (6.6)

1.20 (0.82–1.76)

1.20 (0.82–1.76)

Ante-partum hemorrhage

 No

505 (98.4)

1,521 (98.8)

1.0

1.0

 Yes

8 (1.6)

18 (1.2)

1.34 (0.58–3.10)

1.33 (0.58–3.07)

Premature rupture of membranes

 No

505 (98.4)

1,526 (99.2)

1.0

1.0

 Yes

8 (1.6)

13 (0.8)

1.86 (0.77–4.51)

1.89 (0.77–4.66)

Cephalopelvic disproportion

 No

500 (97.5)

1,464 (95.1)

1.0

1.0

 Yes

13 (2.5)

75 (4.9)

0.51 (0.28–0.92)*

0.50 (0.28–0.92)*

Malpresentation

 No

489 (95.3)

1,447 (94.0)

1.0

1.0

 Yes

24 (4.7)

92 (6.0)

0.77 (0.49–1.22)

0.77 (0.49–1.22)

Placenta previa

 No

507 (98.8)

1,512 (98.2)

1.0

1.0

 Yes

6 (1.2)

27 (1.8)

0.66 (0.27–1.61)

0.65 (0.26–1.61)

Prolonged/obstructed labor

 No

475 (92.6)

1,394 (90.6)

1.0

1.0

 Yes

38 (7.4)

145 (9.4)

0.77 (0.53–1.12)

0.77 (0.54–1.12)

Mode of delivery

 Vaginal

267 (52.0)

818 (53.2)

1.0

1.0

 Elective cesarean

79 (15.4)

154 (10.0)

1.57 (1.16–2.13)**

1.58 (1.16–2.14)**

 Emergency cesarean

167 (32.6)

567 (36.8)

0.90 (0.72–1.13)

0.90 (0.71–1.13)

CI confidence interval

p < 0.05; ** p < 0.01; *** p < 0.001

Prevalence of the perinatal outcomes among cases and controls are shown in Table 2. About one-tenth (10.2%) were low birth weight, 18.7% were delivered preterm and 5.0% had fetal growth restriction. Only seven infants had congenital abnormality, two of whom were AMA offspring. AMA was not associated with any adverse perinatal outcomes after the multivariable analyses. Although compared with YMA mothers, AMA offspring were less likely to be preterm (OR 0.76) or have fetal growth restriction (OR 0.63) the differences were of borderline significance (p = 0.058). We further compared AMA mothers with younger mothers (20–30 years) and still found no significant adverse perinatal outcomes. Rather, AMA was significantly less likely to be associated with fetal growth restriction (OR: 0.55; 95% CI: 0.33–0.90; p = 0.018) in this secondary analysis. Overall, there was no evidence of significant interactions among the variables in logistic regression models nor of poor model calibration as all Hosmer–Lemeshow tests had values of p > 0.05.
Table 2

Perinatal outcomes of delayed childbearing in surviving offspring after multivariable logistic regression

Outcome

Prevalence

Adjusted odds ratio (95% confidence interval)

Cases

Controls

Unmatched

p value

Matched

p value

n = 513 (%)

n = 1539 (%)

Multiple gestation

20 (3.9)

49 (3.2)

0.72 (0.41–1.27)

0.258

0.67 (0.37–1.21)

0.185

Preterm birth (<37 weeks)

96 (18.7)

321 (20.9)

0.80 (0.61–1.05)

0.101

0.76 (0.57–1.01)

0.058

Low birth weight (<2,500 g) [a]

51 (10.2)

177 (11.5)

0.84 (0.59–1.19)

0.316

0.82 (0.57–1.18)

0.292

Fetal growth restriction [b]

25 (5.0)

111 (7.2)

0.63 (0.39–1.01)

0.054

0.63 (0.38–1.02)

0.058

Fetal distress

9 (1.8)

56 (3.6)

0.56 (0.26–1.20)

0.135

0.55 (0.25–1.18)

0.123

1-min Apgar score (0–6) [c]

425 (88.5)

1373 (89.2)

1.08 (0.77–1.53)

0.658

1.08 (0.75–1.55)

0.696

5-min Apgar score (0–6) [d]

126 (26.2)

448 (29.1)

0.93 (0.72–1.19)

0.556

1.00 (0.77–1.30)

0.990

Suspected sepsis

15 (2.9)

57 (3.7)

0.85 (0.46–1.56)

0.592

0.87 (0.46–1.64)

0.656

Hyperbilirubinemia

25 (4.9)

63 (4.1)

1.14 (0.68–1.89)

0.618

1.14 (0.67–1.93)

0.626

SCBU admission

55 (10.7)

200 (13.0)

0.85 (0.61–1.20)

0.362

0.87 (0.61–1.24)

0.436

Exclusive breast feeding

92 (17.9)

279 (18.1)

1.20 (0.88–1.63)

0.247

1.26 (0.85–1.85)

0.246

Missing data: [a] = 12 (0.6%); [b] = 17 (0.8%); [c] = [d] = 33; SCBU = special care baby unit

Adjusted for marital status, maternal occupation, parity, antenatal care, HIV status, cephalopelvic disproportion and mode of delivery

Discussion

Our study suggests that despite the widely-reported “high-risk” attribute of AMA mothers, their surviving offspring may not necessarily be at a greater risk of adverse perinatal outcomes when compared with YMA mothers, even across parity. This finding accords with other studies from both developed [3, 6, 10] and developing [12, 14, 15, 18] countries. For example, a study among 26,795 black women in the USA, reported that the offspring of AMA mothers were not associated with low birth weight and low 5-min Apgar scores [10]. Berkowitz et al. reported that AMA mothers had no increased risk of preterm delivery or of having an infant with fetal growth restriction or low 5-min Apgar scores although a slight elevation in the risk of having a low birth weight infant was observed [3]. Similarly, Hsieh et al. observed that AMA was not associated with low birth weight, low 5-min Apgar scores and admission into intensive care unit but with preterm delivery [12]. Another study in Sri Lanka found that mothers 40 years old and over were not at increased risk of preterm delivery or having an infant with low birth weight or low 5-min Apgar scores as compared to YMA women [14]. One study in Lebanon, found that multiparous mothers (≥40 years) were not at increased risk of having an infant with low birth weight and low 5-min Apgar scores but of intrauterine fetal death as compared to multiparous controls (20–30 years) [15]. Furthermore, one of the limited studies from Nigeria did not find any statistically significant difference between 50 elderly primigravidae (35–40 years) and their younger (20–29 years) controls for low birth weight and low 5-min Apgar scores except for preterm delivery [18]. The favorable outcomes associated with AMA are perhaps attributable to the fact that these mothers are more likely to adopt cautious and healthy lifestyles, prepare themselves psychologically to assume the responsibility of parenthood and take prenatal care more seriously [11]. It is also not unlikely that the predominantly younger AMA cases (<40 years) in this study would have accounted for the lack of significant differences in perinatal outcomes with YMA mothers. However, it is pertinent to observe that studies that have associated AMA independently with one adverse perinatal outcome or more especially in developing countries so far have not been able to establish the precise social, biological and environmental mechanisms for such a relationship [1113, 17].

Generally, the comparability of studies within and across populations on the perinatal outcomes of AMA is usually hindered by variations in the choice of confounders and age classification for AMA (≥35 or ≥40 years) and their controls (≤30 or ≤35 years). For example, AMA was significantly less likely to be associated with fetal growth restriction only when compared with 20–30 years old mothers but not YMA (≤35 years). This would appear to be consistent with studies that reported lack of statistically significant association between AMA and fetal growth restriction [11, 12]. However, in their case–control study of women with IUGR (defined as birthweight ≤10th percentile for gestational age) and non-IUGR, Odibo et al. found AMA to be an independent risk factor for IUGR with the highest odds reported among mothers 40 years or older after adjusting for pregestational diabetes, chronic hypertension, illicit drug use and aneuploid fetuses [25]. Gestational ages were all confirmed by first- or second-trimester ultrasound. Thus the finding in our study would appear inconclusive and warrant further investigation controlling for all potential maternal and fetal confounders as well as ultrasonographic confirmation of gestational age.

The prominent maternal factors associated with AMA in this study are consistent with the widely reported findings in the literature from both developed and developing countries. For example, the association of AMA with increased parity has been reported specifically in developed countries [5, 10] as well as in developing countries [13, 17]. The significantly higher odds (OR: 10.83) associated with grand multiparity in this study as compared to levels typically reported in developed countries would suggest that AMA was less likely to be underpinned by career choices and more likely by poor family planning and cultural predisposition to large family from early age as in many resource-poor countries. The higher odds of AMA among married mothers are consistent with the cultural values in many traditional communities in developing countries which do not encourage single parenting. The association between AMA and maternal occupation may be explained by predominantly anecdotal evidence on the impact of educational and professional/career pursuits on delayed childbearing especially in urban populations. The association of AMA with cesarean delivery is well documented worldwide [1, 2, 12, 13, 15, 17, 19] and in agreement with our earlier finding in this population where the perception of pregnancy related risks was found to be poorer among younger mothers resulting in a higher proportion of emergency cesarean [22]. Studies exploring the association between AMA antenatal care, HIV and cephalopelvic disproportion are relatively sparse. The recognition of the high risk of obstetric complications commonly associated with AMA may account for the significantly higher uptake of antenatal care among affected mothers which was observed with multiparous AMA as compared to multiparous YMA mothers. It is quite plausible that the higher uptake of antenatal care by AMA mothers would have contributed to the higher rates of elective cesarean section and in turn resulted in the minimal risks for adverse perinatal outcomes in this population. However, the uptake of antenatal care by AMA mothers and their controls was significantly less in absolute terms when compared with rates reported in developed countries and thus merits greater attention for improved utilization [26]. It is noteworthy that the perception of pregnancy risk among primiparous AMA and YMA as well as grand-multiparous AMA and YMA was identical as reflected by the lack of statistically significant difference in antenatal care uptake.

The inverse association between AMA and HIV accords with the global epidemiological profile of HIV which shows 45% incidence among the younger adults 25 years or below [27]. The inverse relationship between AMA and cephalopelvic disproportion may be explained by the fact that only one-quarter (25.3%) of AMA mothers were primiparous as compared to over half (53.9%) of YMA mothers. Furthermore, the increased likelihood of multiparous AMA mothers being Muslims as compared to primiparous AMA mothers is corroborated by a similar study from East Asia [13] thus reflecting the more common traditional practice of early marriage among this religious group in developing countries.

While the evidence in the literature on the impact of AMA on perinatal mortality remains conflicting [2, 6, 28], the recruitment criteria for the primary UNHS program precluded an exploration of this association in this study. And like most hospital-based studies in developing countries where majority of deliveries occur outside health facilities, the generalisability of the results still needs to be validated in population-based studies. However, no material threat to the external validity of the main findings for comparable hospital settings is anticipated based on corroborative evidence from other cross-cultural studies. Moreover, the careful selection of controls and systematic matching with cases, the adequately powered sample size, the satisfactory model calibration and narrow confidence intervals are notable strengths of this study. The prospective recruitment of participants into the primary study over a period of approximately 30 months also provided on-going opportunity for timely verification of implausible data commonly resulting from misclassification of patient records, poor maternal recall or lack of clinically validated diagnoses. Overall, this study adds to the current body of evidence demonstrating that high-risk AMA mothers may not necessarily be associated with adverse perinatal outcomes in settings with good antenatal uptake and maternal favorable disposition to planned cesarean section. It also demonstrates the value of context-specific studies to verify the widely reported adverse outcomes attributable to delayed childbearing.

Conflict of interest

None declared by both authors.

Copyright information

© Springer-Verlag 2011