Maternal and perinatal outcomes among nulliparous adolescents in low- and middle-income countries: a multi-country study

Authors


Abstract

Objective

To investigate the risk of adverse pregnancy outcomes and caesarean section among adolescents in low- and middle-income countries.

Design

Secondary analysis using facility-based cross-sectional data from the World Health Organization (WHO) Global Survey on Maternal and Perinatal Health.

Setting

Twenty-three countries in Africa, Latin America, and Asia.

Population

Women admitted for delivery in 363 health facilities during 2–3 months between 2004 and 2008.

Methods

We constructed multilevel logistic regression models to estimate the effect of young maternal age on risks of adverse pregnancy outcomes.

Main outcome measures

Risk of adverse pregnancy outcomes among young mothers.

Results

A total of 78 646 nulliparous mothers aged ≤24 years and their singleton infants were included in the analysis. Compared with mothers aged 20–24 years, adolescents aged 16–19 years had a significantly lower risk of caesarean section (adjusted OR 0.75, 95% CI 0.71–0.79). When the analysis was restricted to caesarean section indicated for presumed cephalopelvic disproportion, the risk of caesarean section was significantly higher among mothers aged ≤15 years (aOR 1.27, 95% CI 1.07–1.49) than among those aged 20–24 years. Higher risks of low birthweight and preterm birth were found among adolescents aged 16–19 years (aOR 1.10, 95% CI 1.03–1.17; aOR 1.16, 95% CI 1.09–1.23, respectively) and ≤15 years (aOR 1.33, 95% CI 1.14–1.54; aOR 1.56, 95% CI 1.35–1.80, respectively).

Conclusions

Adolescent girls experiencing pregnancy at a very young age (i.e. <16 years) have an increased risk of adverse pregnancy outcomes.

Introduction

Pregnancy and childbirth are the leading causes of death in adolescent girls in low- and middle-income countries,[1] and obstetric risk is highest by far in sub-Saharan Africa.[2] It is estimated that about 15 million adolescents aged 15–19 years give birth each year, and more than 90% of these births occur in low- and middle-income countries.[1] In 2007, the number of births per 1000 women aged 15–19 years was as high as 121 in sub-Saharan Africa, and 74 in Latin America and the Caribbean, and in many parts of the world the decline in adolescent birth rates has stagnated since 2000.[3]

Pregnancy during adolescence has been associated with higher risks of adverse pregnancy outcomes such as low birthweight, preterm delivery, and perinatal and maternal death.[4, 5] However, pregnant adolescents are more likely to be socially deprived: they are more likely to be poorer, less educated, and single, and to have fewer antenatal care visits than older mothers.[4-6] Hence, the contribution of young age itself to the risk of adverse pregnancy outcomes among adolescents is difficult to estimate, as the risk may be at least partly attributable to the unfavourable sociodemographic characteristics of adolescent mothers.[7]

Despite the fact that most adolescent pregnancies occur in low- and middle-income countries, large studies from such countries are scarce, and most of these studies have not taken into account the confounding effect of sociodemographic factors.[8-10] We therefore conducted an analysis of data from the Global Survey on Maternal and Perinatal Health, which was a large multi-country study carried out in 24 mainly low- and middle-income countries in Africa, Latin America, and Asia, to assess the effect of young maternal age on pregnancy outcomes after adjustment for sociodemographic characteristics.

Methods

Study design

A secondary analysis of data from the World Health Organization (WHO) Global Survey on Maternal and Perinatal Health was conducted. Methodological details of this survey have been published elsewhere.[11, 12] Briefly, the survey was a multi-country, facility-based, cross-sectional study. A stratified multistage cluster sampling design was used to obtain a sample of countries and health facilities worldwide. The survey was implemented in 24 countries and 373 health facilities. In each country, the capital city was included in the sample. In addition, two provinces were randomly selected from other administrative areas. The countries included in the survey were from three continents: Africa (Algeria, Angola, Democratic Republic of Congo, Niger, Nigeria, Kenya, and Uganda), Latin America (Argentina, Brazil, Cuba, Ecuador, Mexico, Nicaragua, Paraguay, and Peru), and Asia (Cambodia, China, India, Japan, Nepal, the Philippines, Sri Lanka, Thailand, and Vietnam). Only health institutions that reported at least 1000 deliveries per year and were capable of performing caesarean sections were eligible to participate. A random sample of up to seven institutions from each country was selected. All women admitted for delivery during a fixed recruitment period were included in the survey. The period of recruitment (data collection) was arbitrary, defined as 3 months for institutions with up to 6000 deliveries per year, and 2 months for those with more than 6000 deliveries. The survey was conducted between 2004 and 2005 in Africa and Latin America, and between 2007 and 2008 in Asia.

Data collection

Data were collected for institutions and individuals. Data for individuals were directly extracted from the patient records by trained medical staff before maternal hospital discharge or within 7 days of delivery. Maternal age referred to age in completed years at the time of delivery. Maternal weight in Africa and Latin America was measured at the last antenatal care visit, whereas in Asia it was measured upon maternal admission for delivery. Gestational age at birth was defined as the number of completed weeks of gestation based on the estimated delivery date in the clinical records. Consistent with previous studies,[13, 14] the operational definition of caesarean section indicated for cephalopelvic disproportion (CPD) in our study was based on presumed CPD, and included caesarean sections indicated for dystocia, failure to progress, and failed vacuum extraction or forceps delivery, but excluded those indicated for fetal distress, breech and other types of malpresentation, and failed induction. Maternal and early neonatal deaths were defined as intra-hospital deaths that occurred on or before the eighth day postpartum and seventh day after delivery, respectively. Stillbirths included both macerated and fresh stillbirths.

Institutional data were collected on a standard form by the hospital coordinators, in consultation with the director or head of obstetrics. Data on the provision of maternal and perinatal care were obtained, including: laboratory tests; anaesthesiology resources; services provided for intrapartum care, delivery, and care of the newborn baby; and the presence or absence of basic emergency medical and obstetric care facilities, intensive care units, and human training resources.

Statistical analysis

The study population was restricted to nulliparous women because of the small number of multiparous younger mothers and the strong correlation between parity and maternal age. Only mothers who had a singleton neonate of birthweight ≥500 g or gestational age ≥22 weeks – if the birthweight was missing were included in the analysis, as a birthweight or gestational age outside this range is considered a miscarriage.[15] Age was categorised as ≤15, 16–19, and 20–24 years, and, similar to previous studies,[4, 8, 16, 17] those over 25 years of age were excluded from the analysis. Japan was excluded from the analysis as our aim was to focus on low- and middle-income countries.

Region-wise descriptive analyses were conducted to account for the heterogeneity between regions, as the quality of medical services and the social and cultural backgrounds of adolescent mothers were expected to differ by region. The region-wise characteristics of mothers in the three age groups were compared using chi-square tests.

To estimate the effect of young maternal age on the risks of caesarean section, caesarean section indicated for CPD, low birthweight (<2500 g), preterm delivery (<37 completed weeks of gestation), and perinatal death (early intra-hospital death or stillbirth), we constructed multilevel logistic regression models with random effects for facilities and countries. The significance of the random effects was determined using a likelihood ratio test by comparing models nested within models with additional levels.[18] The variability of outcomes was assessed, and caterpillar plots of the random-effect estimates for each country are shown for low birthweight only (Figure 1). For the analyses, four models are presented. In order to understand the independent roles of biological versus environmental factors in the association between adolescent pregnancy and adverse birth outcomes, adjustment was made for country and facility variables in the first model, and individual variables such as maternal education, marital status, and antenatal care visits in the second. For analyses on caesarean section, caesarean section indicated for CPD, low birthweight, and perinatal mortality, further adjustment was made for maternal height (short maternal height being defined as <1.50 m) and gestational age at birth in the third and fourth models. Individual- and facility-level variables were available in the survey data, whereas country-level variables were obtained from the World Bank Development Indicators and Lancet series on maternal and child mortality.[19] Country variables included in the model were gross national income (GNI) per capita, classified according to the World Bank categorisation as low ($935 or less), lower–middle ($936–3705) and upper–middle ($3706–11 455), maternal mortality ratio (MMR) per 100 000 live births, and early neonatal mortality (ENM) per 1000 live births. The facility capacity index score was considered as a facility-level variable, and was used to reflect the level of services available in each of the facilities and to summarise the capacity of an institution to provide obstetric care.[20] The analyses were not adjusted for maternal body mass index (BMI) because of variation in the timing of maternal weight measurements among countries. Data on maternal morbidities such as eclampsia, diabetes, severe anaemia (defined as haemoglobin, HB < 7 g/dl), sexually transmitted infections, and HIV infection were available, and the multiple regression analyses were only adjusted for those associated with maternal age in the univariate analysis, in order to restrict the number of confounding factors included in the final models. For multilevel logistic regression analyses on caesarean section and caesarean section indicated for CPD, facilities with no provision for caesarean section delivery, and countries with > 40% missing data on maternal height (Angola, 42.3%; Kenya, 87.5%; Brazil, 68.1%), were excluded. Statistical analysis was performed in stata/MP 12.0 (Stata Corp. LP, College Station, TX, USA), and P < 0.05 was considered to be statistically significant.

Figure 1.

Random-effect estimations for low birthweight: country-level effects in rank order, with approximate 95% confidence intervals. DRC, Democratic Republic of Congo.

Results

The WHO Global Survey on Maternal and Perinatal Health collected data on a total of 290 610 deliveries. A total of 119 551 mothers and their infants were retained in the analysis after exclusion of data from Japan, multiparous births, multiple births, and infants with birthweight <500 g or gestational age <22 weeks, if birthweight was missing. Table 1 presents the distribution of women by age group. The overall proportion of adolescent mothers (i.e. ≤19 years) in the study population was 35.1, 35.5, and 11.5% in Africa, Latin America, and Asia, respectively. When restricted to mothers ≤24 years of age, the sample size was 78 646, with 2361 (3%), 27 398 (34.8%), and 48 887 (62.2%) mothers aged ≤15, 16–19, and 20–24 years, respectively.

Table 1. Proportion of mothers by age group
RegionCountryTotal (n)Maternal age at delivery (years)
≤15 (n [%])16–19 (n [%])20–24 (n [%])≥25 (n [%])
  1. DRC, Democratic Republic of Congo.

  2. Numbers shown are for nulliparous mothers who had singleton infants with a birthweight ≥500 g, or a gestational age of ≥22 weeks if birthweight was missing.

AfricaAlgeria50133 (0.1)201 (4.0)1671 (33.3)3138 (62.6)
Angola1860102 (5.5)1197 (64.4)437 (23.5)124 (6.7)
DRC249081 (3.3)978 (39.3)897 (36.0)534 (21.5)
Kenya8540134 (1.6)2887 (33.8)4049 (47.4)1470 (17.2)
Niger200648 (2.4)933 (46.5)727 (36.2)298 (14.9)
Nigeria245393 (3.8)561 (22.9)647 (26.4)1152 (47.0)
Uganda484390 (1.9)2241 (46.3)1814 (37.5)698 (14.4)
Latin AmericaArgentina4304127 (3.0)937 (21.8)1214 (28.2)2026 (47.1)
Brazil6535385 (5.9)2285 (35.0)2354 (36.0)1511 (23.1)
Cuba6193147 (2.4)1503 (24.3)2089 (33.7)2454 (39.6)
Ecuador4649272 (5.9)1933 (41.6)1535 (33.0)909 (19.6)
Mexico7743263 (3.4)2499 (32.3)2843 (36.7)2138 (27.6)
Nicaragua2340204 (8.7)1133 (48.4)715 (30.6)288 (12.3)
Paraguay129932 (2.5)368 (28.3)609 (46.9)290 (22.3)
Peru6889171 (2.5)1924 (27.9)2679 (38.9)2115 (30.7)
AsiaCambodia29161 (0.03)266 (9.1)1398 (47.9)1251 (42.9)
China92390 (0.0)455 (4.9)4140 (44.8)4644 (50.3)
India112172 (0.02)727 (6.5)7640 (68.1)2848 (25.4)
Nepal491014 (0.3)993 (20.2)2930 (59.7)973 (19.8)
Philippines584396 (1.6)1450 (24.8)2367 (40.5)1930 (33.0)
Sri Lanka670210 (0.2)835 (12.5)2191 (32.7)3666 (54.7)
Thailand465182 (1.8)863 (18.6)1378 (29.6)2328 (50.1)
Vietnam69164 (0.1)229 (3.3)2563 (37.1)4120 (59.6)

Maternal and infant characteristics by age group are presented in Table 2. Adolescent mothers were more likely to have fewer antenatal care visits, to have a lower BMI, to be of shorter stature, to be single, and to have low-birthweight infants, and were less likely to have a caesarean delivery in all three regions, and to have less education in Africa and Latin America. Risks of maternal and perinatal deaths were significantly higher among adolescent mothers in Africa, compared with non-adolescents mothers.

Table 2. Maternal and infant characteristics by age group
CharacteristicsAge group (years)
AfricaLatin AmericaAsia
≤1516–1920–24≤1516–1920–24≤1516–1920–24
  1. Africa includes Algeria, Angola, Democratic Republic of Congo, Kenya, Niger, Nigeria, and Uganda. Latin America includes Argentina, Brazil, Cuba, Ecuador, Mexico, Nicaragua, Paraguay, and Peru. Asia includes Cambodia, China, India, Nepal, the Philippines, Sri Lanka, Thailand, and Vietnam.

  2. Data are the region-wise prevalence of that characteristic in that age group; NR, not reported.

  3. Less than 5% of region-wise data were missing for all variables, with the exception of BMI (Africa, 42%; Latin America, 18%), height (Africa, 38%; Latin America, 17%), number of antenatal care visits (Africa, 20%), and maternal education (Africa, 14%).

  4. *P < 0.05 in Africa, **Latin America, and ***Asia by the chi-square test (two degrees of freedom [df], except for antenatal care visits [4 df]).

  5. ****CPD includes dystocia, failure to progress, and failed vacuum extraction or forceps delivery.

  6. *****Perinatal death includes fresh and macerated stillbirths and early neonatal death, which was defined as the intra-hospital death of a liveborn neonate during the first 7 days after delivery or earlier if the discharge occurred before 7 days.

  7. ******Maternal death refers to intra-hospital deaths that occurred on or before the 8th day postpartum.

Marital status: single (%) * , ** , *** 53.137.117.949.037.428.353.421.66.2
Maternal education: none (%) * , ** 26.916.78.21.91.10.82.98.38.4
Maternal BMI <20 kg/m2 (%) * , ** , *** 12.48.54.22.82.71.77.77.75.0
Maternal height <1.50 m (%) * , ** , *** 26.211.45.819.713.711.019.218.014.7
Number of antenatal care visits (%) * , ** , ***
019.010.54.67.04.43.07.16.24.6
1–346.848.635.919.013.810.039.131.231.9
≥434.240.959.574.181.887.053.862.563.6
Caesarean section (%) * , ** , *** 11.49.111.631.127.033.413.414.122.4
Caesarean section indicated for CPD **** (%) ** , *** 6.75.15.414.410.412.07.64.16.8
Low birthweight, <2500 g (%) * , ** , *** 19.712.39.613.39.58.321.217.914.5
Preterm birth, <37 weeks (%) * , ** 21.815.811.012.18.17.214.912.311.5
Perinatal death ***** (%) * 7.65.04.51.81.31.21.42.52.3
Maternal death (per 10 000) ****** , * 73.129.119.66.30.82.9NR10.39.4

Table 3 demonstrates the association between maternal age and the risk of all caesarean sections and caesarean section indicated for CPD. After adjustment for country- and facility-level effects and covariates (model 1), the risk of all caesarean sections was significantly lower among adolescent mothers aged 16–19 years (adjusted odds ratio, aOR 0.74; 95% confidence interval, 95% CI 0.73–0.78), whereas it was not significantly different among mothers aged ≤ 15 years (aOR 1.05, 95% CI 0.93–1.19) compared with non-adolescent mothers. With further adjustment for sociodemographic characteristics and number of antenatal care visits (model 2), and maternal height (model 3), the lower risk of caesarean section among older adolescent mothers remained statistically significant. For caesarean section indicated for CPD, those aged ≤15 years had a significantly (1.32 times) higher risk (95% CI 1.13–1.55) compared with mothers aged 20–24 years, whereas older adolescents had a significantly lower risk (aOR 0.82, 95% CI 0.77–0.88) after adjustment for country- and facility-level effects and covariates (model 1). These associations remained similar even after adjustment for sociodemographic characteristics, number of antenatal care visits (model 2) and maternal height (model 3).

Table 3. Risk of all caesarean sections and caesarean section indicated for cephalopelvic disproportion among adolescent mothers, compared with mothers aged 20–24 years
OutcomesMaternal age (years)Model 1Model 2Model 3
aOR95% CIaOR95% CIaOR95% CI
  1. a

    Cephalopelvic disproportion includes dystocia, failure to progress, and failed vacuum extraction or forceps delivery.

  2. Three-level structure logistic random effects regression models were used to obtain the ORs: individual (level 1); facility (level 2); and country (level 3).

  3. Facilities with no provision for caesarean section delivery and countries with >40% missing data for maternal height (Angola, 42.3%; Kenya, 87.5%; Brazil, 68.1%) were excluded.

  4. Model 1: adjusted for facility complexity score at the facility level, and gross national income per capita and maternal mortality rate at the country level.

  5. Model 2: model 1, with further adjustment for marital status, maternal education, and number of antenatal care visits at the individual level.

  6. Model 3: model 2, with further adjustment for maternal height at the individual level.

All caesarean sections
≤151.05(0.93–1.19)1.09(0.97–1.24)1.05(0.93–1.20)
16–190.74(0.73–0.78)0.76(0.72–0.80)0.75(0.71–1.79)
Caesarean section indicated for cephalopelvic disproportion a
≤151.32(1.13–1.55)1.34(1.14–1.59)1.27(1.07–1.49)
16–190.82(0.77–0.88)0.84(0.78–0.91)0.82(0.76–0.88)

Table 4 illustrates the associations between maternal age and low birthweight, preterm delivery, and perinatal death. After adjustment for country- and facility-level effects and covariates (model 1), there were significant dose-dependent relationships between maternal age and low birthweight and preterm delivery, with younger mothers having higher risks of delivering low-birthweight (aOR 1.71, 95% CI 1.51–1.93) or preterm (aOR 1.74, 95% CI 1.53–1.99) babies. Although the increased risks of delivering low-birthweight and preterm infants among all adolescent mothers were slightly decreased after further adjustment for sociodemographic characteristics and number of antenatal care visits (model 2), and gestational age at birth (model 4), the risks still remained significantly higher than in non-adolescent mothers. Risk of perinatal death was significantly higher in younger (aOR 1.53, 95% CI 1.19–1.96) and older (aOR 1.15, 95% CI 1.03–1.27) adolescent pregnancies than in non-adolescent pregnancy with adjustment for country- and facility-level effects and covariates (model 1). Further adjustments made for sociodemographic characteristics and the number of antenatal care visits (model 2) and gestational age at birth (model 4) attenuated the association between young maternal age and perinatal death.

Table 4. Risks of low birthweight, preterm delivery, and perinatal death among adolescent mothers, compared with mothers aged 20–24 years
OutcomesMaternal age (years)Model 1Model 2Model 4
aOR95% CIaOR95% CIaOR95% CI
  1. Three-level structure logistic random effects regression models were used to obtain the ORs: individual (level 1); facility (level 2); and country (level 3).

  2. a

    Perinatal death includes fresh and macerated stillbirths and early neonatal death, which was defined as the intra-hospital death of a liveborn neonate during the first 7 days after delivery, or earlier if the discharge occurred before 7 days.

  3. Model 1: adjusted for facility complexity score at the facility level, and gross national income per capita and maternal mortality rate at the country level.

  4. Model 2: model 1, with further adjustment for marital status, maternal education, and number of antenatal care visits at the individual level.

  5. Model 4: model 2, with further adjustment for gestational age at birth at the individual level.

Preterm birth (<37 weeks of gestation)
≤151.74(1.53–1.99)1.56(1.35–1.80)
16–191.23(1.16–1.30)1.16(1.09–1.23)
Low birthweight (<2500 g)
≤151.71(1.51–1.93)1.53(1.34–1.74)1.33(1.14–1.54)
16–191.22(1.15–1.28)1.16(1.10–1.23)1.10(1.03–1.17)
Perinatal death a
≤151.53(1.19–1.96)1.20(0.91–1.59)0.99(0.87–1.12)
16–191.15(1.03–1.27)1.05(0.93–1.18)0.96(0.70–1.31)

The results for adverse perinatal outcomes in model 4 were not altered after additional adjustment for eclampsia and malaria (data not shown).

Discussion

Main findings

Using a large multicountry data set, we investigated the delivery outcomes of adolescent mothers in 23 low- and middle-income countries. After adjusting for country- and facility-level effects, we found that adolescent pregnancy was associated with increased risks of low birthweight, preterm delivery, and perinatal mortality, with a general tendency for poorer outcomes in younger adolescents. After further adjustment for confounding factors at the individual level, only the increased risk of perinatal mortality was attenuated. After adjustment for facility- and country-level effects, and potential confounding factors at the individual level, the risks of all caesarean sections and caesarean section with indication for CPD were significantly lower among adolescents aged 16–19 years than among women aged 20–24 years. Higher risks were observed among younger adolescents aged ≤15 years than among women aged 20–24 years, but only the risk of caesarean section with indication for CPD reached statistical significance. A higher rate of intra-hospital maternal death was observed among younger mothers (those aged ≤19 years) in Africa.

Limitations and strengths

To the best of our knowledge, this is the largest cross-regional study to report on the pregnancy outcomes of adolescents from low- and middle-income countries. We adjusted for country and health facility effects in addition to sociodemographic characteristics, without which valid conclusions could not have been drawn as to whether the effects observed were attributable to biological mechanisms or environmental factors.[6] Most previous studies in low- and middle-income countries in particular have failed to adjust for sociodemographic factors,[8-10] with a few exceptions.[4, 21]

This study has several limitations. First, the health facilities that implemented the study had an annual delivery rate >1000, and were located mainly in urban settings, and thus the results are not generalisable to the broader population, where the majority of adolescents tend to deliver outside of facilities or at smaller hospitals. Secondly, residual confounding probably exists in our analyses. We attempted an adjustment for the sociodemographic factors available in this survey, but robust indicators of socio-economic status, such as family income and employment, were not available. As a result, the extent to which the effect of young maternal age is confounded with poverty and other forms of social marginalisation and vulnerability could not be assessed in this study. Moreover, we did not have any information on smoking, gestational weight gain, and sexually transmitted diseases, which are risk factors for low birthweight and preterm delivery.[22-24] Also, our definition of CPD was based on presumptive CPD, which included dystocia, failure to progress, and failed vacuum extraction or forceps delivery. Thus, data should be interpreted with caution as there could have been some degree of misclassification; however, in areas with limited diagnostic facilities, such as most of the areas in which our study was conducted, the diagnosis of CPD cannot be easily performed, and thus we believe that the use of this presumptive definition can be justified.

Interpretation

We found that adolescent pregnancy was independently associated with increased risks of low birthweight and preterm delivery, consistent with previous studies in developed and low- and middle-income countries.[4, 7, 16, 25] As for proposed biological mechanisms, adolescent mothers who are themselves still developing may compete with their fetus for nutrients, resulting in low birthweights.[7] Alternatively, the gynaecological immaturity of adolescents, which is characterised by a short uterine cervix and alkaline vaginal pH, may predispose them to an increased risk of subclinical lower genital infection, leading to an increased risk of preterm delivery.[26]

The increased risk of perinatal death among infants born to adolescent mothers disappeared after further adjustment for sociodemographic characteristics, the number of antenatal care visits, and gestational age at birth. This finding supports the results of previous studies, which have concluded that the increased risk of neonatal mortality among adolescents is largely attributable to preterm birth and the socio-economic circumstances of younger mothers.[4, 16, 21, 27, 28]

Consistent with previous studies, we found a significantly lower risk of caesarean section deliveries among adolescents aged 16–19 years compared with women aged 20–24 years.[4, 29-31] This may be partly explained by the fact that adolescents generally have smaller babies than older women.[30] The lower rates of caesarean section among those aged 16–19 years may not necessarily indicate that younger mothers had fewer obstetric complications, as the decision to perform a caesarean section is likely to be influenced by the practitioner's policy and maternal opinion.[32, 33]

We found a significantly higher risk of caesarean section with indication for CPD among adolescent mothers aged ≤15 years, which was consistent with previous studies.[8, 10, 34] As adolescents are still in the growing phase, the pelvis has not yet reached its maximum size, and is expected to be a cause of CPD or obstructed labour.[34] Furthermore, the onset of menarche is known to occur later in low- and middle-income countries,[5] and shorter intervals between menarche and pregnancy (low gynaecological age) are associated with the risk of having an inadequate pelvis size or an immature birth canal.[5, 35] A significantly higher rate of maternal death was observed among adolescent mothers (those aged <20 years) in Africa. Although concrete conclusions may not be drawn because of the lack of adjustment for potential confounding factors, the fact that the youngest mothers were at a higher risk of caesarean section indicated for CPD implies that there may be a causal relationship, which can at least in part be explained by obstructed labour.

Conclusion

In conclusion, our results demonstrate that adolescent pregnancy occurring among the very young (i.e. ≤15 years old) is associated with an increased risk of caesarean section indicated for CPD, and that younger age increases the risk of low birthweight and preterm delivery. Social deprivation in terms of sociodemographic characteristics, less antenatal care, and preterm birth among adolescents contributed greatly to the increased risk of perinatal death. These findings emphasise the importance of appropriate multidisciplinary interventions to prevent early pregnancies and to provide antenatal and obstetric care of adolescent mothers in order to minimise their socio-economic deprivation and risk of adverse birth outcomes.

Disclosure of interests

None of the authors has any conflict of interests to disclose.

Contribution to authorship

GT, RM, SG, and AK analysed the data and wrote the article. MRT, ARB, JV, AC, and JPS designed the secondary analysis, contributed to data collection, and provided advice. OE, KS, MRT, JV, and JPS contributed to editing the article. All authors read and approved the final version of the article.

Details of ethics approval

Written consent was obtained from all ministries of health of the participating countries and from the directors of the selected facilities. Individual informed consent was not obtained because this study was a cluster-level study in which the data were collected from medical records without any individual identification. The WHO Ethics Review Committee and that of each country independently approved the protocol.

Funding

This study was financially supported by: the UNDP/UNFPA/WHO/World Bank Special Program of Research, Development and Research Training in Human Reproduction (HRP); the Department of Reproductive Health and Research (RHR); the Ministry of Health, Labour and Welfare of Japan; the United States Agency for International Development (USAID); the Ministry of Public Health, Beijing, China; and the Indian Council of Medical Research, Delhi, India. GT, AK, and KS were supported partly by a research grant from the Ministry of Health, Labour and Welfare of Japan (Global Issues Research Grant, no. 09151353).

Acknowledgements

The authors thank all those who contributed to the study design and implementation, including the researchers, study coordinators, data collectors, data clerks, and staff of the Ministries of Health and WHO offices, and the participants in the study.

Commentary on ‘Maternal and perinatal outcomes among nulliparous adolescents in low- and middle-income countries: a multi-country study’

This article by Ganchimeg et al. in this issue of BJOG makes interesting reading. It investigated the risks of caesarean section and adverse pregnancy outcomes among adolescents in low- and middle-income countries. Data were obtained from the WHO Global Survey on Maternal and Perinatal Health for singleton births to 29 759 and 48 887 nulliparous mothers, aged ≤19 and 20–24 years, respectively, in 363 healthcare facilities in 23 countries in Africa, Latin America, and Asia. The results show that compared with mothers aged 20–24 years, adolescents aged 16–19 years had lower risks of caesarean section, whereas mothers aged less than 15 years had higher risks for caesarean section arising from cephalopelvic disproportion. Adolescents also had higher risks of low birthweight babies and preterm births.

These findings confirm what has been reported in the literature (Harrison BJOG 1985;92:1–119; Kurth et al. PLoS One 2010;5:e14367); however, this represents one of the largest data sets on adolescent births ever reported, and the fact that it covered low- and middle-income countries across three continents makes the findings even more profound. The research procedure is also internally valid, with processes, protocols, and data collection procedures adequately reported.

Despite the strengths of the article, several concerns remain. First, the data were collected between 2004 and 2005 in Africa and Latin America, and between 2007 and 2008 in Asia, and are only being reported in 2013. It is not known whether the determinants of adolescent pregnancy first identified for the cohorts of adolescents at the time of data collection are still relevant today. Secondly, the study was conducted for leading health institutions in major cities, especially in public sector facilities with delivery rates of >1000 per year. By contrast, deliveries in smaller hospitals as well as those occurring outside the healthcare system were not included. In many low- and middle-income countries, especially in sub-Saharan Africa, up to 50% of women and adolescents may deliver outside the healthcare system. The non-inclusion of adolescents who deliver in different healthcare facilities as well as those delivering outside the healthcare system limits the external validity of the study.

Thirdly, although the study reported higher risks of perinatal deaths among adolescents, it was unable to explore the risks for maternal mortality in adolescent births because of the small numbers in the subgroups. Therefore, there remains the unresolved question as to whether adolescent births in low- and middle-income countries are associated with higher risks of maternal mortality. Also unexplored is the association between adolescent pregnancies and the risks of severe obstetric morbidity, such as near-miss maternal deaths, eclampsia, vesicovaginal fistulae, and anaemia, which have been widely reported in the literature. Although this study is highly relevant, a community-based study that assesses birth outcomes in adolescents delivering in all settings is needed to accurately quantify the risks associated with adolescent pregnancy in low- and middle-income countries. This may encourage a wider debate on community-based versus hospital-based care for adolescents in such countries.

Disclosure of interests

F.O. was a reviewer of the initial article.

  • F Okonofua

  • Professor of Obstetrics and Gynaecology, University of Benin, Benin City, Nigeria

Ancillary