Prevalence, risk factors and adverse outcomes of intrapartum related asphyxia amongst newborns from mothers with low-risk deliveries in peri-urban Gambian health facilities: a retrospective cohort study

Abstract

Introduction

Despite significant progress in perinatal care, intrapartum-related asphyxia remains a leading cause of neonatal morbidity and mortality, especially in the poorest regions of the world such as sub-Saharan Africa. The objective of this study is to determine the prevalence, risk factors, and adverse outcomes of intrapartum-related asphyxia among low-risk health facility-based deliveries in The Gambia.

Methods

This is a secondary analysis of a randomised clinical trial conducted in western Gambia between October 2017 and May 2021. Women with low-risk pregnancies in labour and their live neonates were recruited at two public health facilities. The prevalence, risk factors, and neonatal outcomes of intrapartum-related asphyxia, defined as 1-min Apgar score < 7 were determined using descriptive methods, and unadjusted and adjusted logistic regression.

Results

Among 6,758 neonates, 257 (3.8%) experienced intrapartum-related asphyxia, 93 (36.2%) of them required hospitalisation. The neonatal case fatality rate was 13.9% (34/245) among newborns with intrapartum-related asphyxia. Neonatal death (OR = 19.56, 95% CI:12.02–31.55, p-value < 0.001) and hospitalisation (OR = 11.91, 95% CI:8.90–15.90, p-value < 0.001) were significantly higher among these neonates. Severe intrapartum-related asphyxia cases had a significantly high risk for neonatal death (aOR:71.8, 95%CI, 36.00–140.40, p-value < 0.001) and hospitalisation (aOR:19.5, 95%CI, 10.80–35.00, p-value < 0.001); however, this risk remained considerable even in mild/moderate cases. Adjusted analysis revealed macrosomia as a major modifiable risk factor (aOR: 3.31, 95%CI, 1.78–6.15, p-value < 0.001), and several identifiable risk factors such as history of previous stillbirths (aOR: 2.79, 95%CI, 1.41–5.52, p-value = 0.003) or miscarriage (aOR: 1.65, 95%CI, 1.08–2.52, p-value = 0.02), and primiparity (aOR: 1.69, 95%CI, 1.16–2.46, p-value < 0.001).

Conclusions

Intrapartum-related asphyxia is a major contributor to neonatal morbidity and mortality among low-risk pregnant women in peri-urban Gambia. A significant proportion of intrapartum-related asphyxia is preventable with timely identification of risk factors by improved antenatal and obstetric care followed by appropriate intrapartum management. Availability of adequate equipment, infrastructures, and skilled staff to help neonates breathe at birth are urgently required to reach the Sustainable Development Goal 3.2 of reducing by 2030 neonatal mortality to 12 deaths per 1000 live births.

Trial registration

NCT03199547: Clinicaltrials.gov. Registered on 23rd June 2017. DOI: https://dx.doi.org/10.18203/2349-3259.ijct20220110.

Background

In 2021, 2.3 million neonates died worldwide, representing approximately half of all under-five deaths [1]. Intrapartum-related asphyxia (IRA), also known as birth asphyxia remains a leading cause of neonatal mortality, contributing to 24% of neonatal deaths worldwide [1]. In addition, IRA leads to short- and long-term morbidities, including early hospitalisation, neurodevelopment and motor delays, seizures, or cerebral palsy [2]. According to the World Health Organization, in 2019 IRA accounted for 61 million disability-adjusted life years (DALYs) [3]. In sub-Saharan Africa, the region with the highest burden of neonatal mortality, IRA was the leading cause (30%) of overall neonatal deaths by day 28 in 2019, 60% of which occurred during the first week of life [1].

The main risk factors associated with IRA can be categorised as maternal, pregnancy-related, or neonatal. Maternal factors include socio-demographic characteristics such as age, educational status, and marital status [4]. In addition, pregnancy-related conditions such as primiparity, severe maternal hypotension or hypertension/pre-eclampsia, and antepartum haemorrhage [5] are important in the causal pathway for IRA. Labour complications such as placental abruption or umbilical cord prolapse are major intrapartum sentinel events leading to IRA, with meconium-stained amniotic fluid a physical manifestation of foetal distress [6]. In sub-Saharan Africa, additional neonatal risk factors such as birth weight (low or macrosomia), multiple gestations, and preterm delivery [7, 8] have been previously identified.

IRA can be identified by three different types of measurements: (i) process-based indicators (i.e., prolonged labour or foetal distress detected via partograph or cardiotocography); (ii) clinical sign-based indicators (i.e., low Apgar scores, foetal or neonatal acidosis), and (iii) outcome-based indicators (i.e., stillbirth or neonatal mortality, signs of neonatal encephalopathy) [9]. In sub-Saharan Africa and other low-income regions, the Apgar score remains routinely and widely used to detect IRA. The Apgar score is a quick assessment performed on a neonate after birth and comprises the following five clinical signs: heart rate, respiratory effort, muscle tone, reflex irritability, and colour. Each sign is assigned a value of 0, 1, or 2, with the sum comprising the score out of 10 [10]. The Apgar score is reported at 1-min and 5-min after birth for all neonates, and at 5-min intervals thereafter until 20 min for infants with a score less than 7 [10]. A 1-min Apgar score provides an indication of intrapartum health and the neonatal response to the process of birth [11, 12]. However, Apgar scores can be subjective in the absence of robust training and prone to intra- and inter-observer variability [13]. With limited training, values of Apgar scores are often over-estimated thus underestimating the burden of IRA in resource-constrained settings as only the more severe cases are identified. Despite these limitations, understanding the factors associated with a low 1-min Apgar score may help to improve the transition from intrauterine to extrauterine life [14], enable life-saving resuscitation for those at highest risk, and eventually decrease the mid and long-term sequelae from IRA.

Evidence on risk factors for IRA in West Africa is limited, despite its contribution towards adverse neonatal outcomes in this region of high neonatal mortality. In the Gambia, a small country in the West African region, the neonatal mortality rate (NMR) in 2020 was estimated at 26 deaths per 1000 live births [15], more than twofold higher than the target of 12 deaths per 1000 live births set in the Sustainable Development Goal (SDG) for 2030. In a cohort study that included 829 deliveries at a Gambian urban health facility, IRA accounted for 31% of neonatal deaths, representing the second most common diagnosis [16].

The study presented here aims to determine the prevalence and the risk factors associated with IRA as detected by the 1-min Apgar score, in a large cohort of newborns at two Gambian peri-urban health facilities. Detailed and context-specific understanding of the risk factors associated with IRA has the potential to improve perinatal care and optimise the transition to extra-uterine life [17].

Methods

Study design

This is a secondary analysis of the PregnAnZI-2 trial [18] a phase III, double-blind, placebo-controlled individually randomised clinical trial in which ~ 12,000 women in labour in The Gambia and Burkina Faso received either azithromycin or placebo (ratio 1:1) from October 2017 to May 2021 [18]. The trial showed no differences in IRA by trial arm [19] and therefore we have included the full Gambian cohort in this retrospective cohort study. A separate analysis is conducted for the Burkina Faso cohort.

Study setting

The climate of The Gambia is typical of the sub-Sahel region, with a long dry season from November to May and a short rainy season between June and October. The population of the catchment area is representative of The Gambia and includes all the main ethnic groups. Illiteracy rates are high, up to 55% among women [18].

The study was conducted at two Government-run health facilities situated in western Gambia, namely Serekunda Health Centre (SHC) and Bundung Maternal and Child Health Hospital (BMCHH) [18]. These two health facilities could not consistently provide obstetric ultrasound scans during antenatal visits throughout our study period. At the time of data collection, SHC could not provide emergency obstetric care, including surgical care; and lacked a neonatology unit. Thus, women in need of an emergency caesarean section were referred to the Kanifing General Hospital, 3.4 km away from SHC. Neonates born at SHC and requiring hospital care were transferred to the neonatal unit at Kanifing General Hospital or to BMCHH. BMCHH provided emergency obstetric and neonatal care and on rare occasions transferred women or neonates to the national referral centre, the Edward Francis Small Teaching Hospital in Banjul, about 15.5 km from BMCHH.

Participants

The PregnAnZI-2 trial recruited pregnant women aged ≥ 16 years receiving antenatal, and intrapartum care at the study health facilities. Women were excluded if they had known HIV infection, any chronic or acute conditions, planned (elective) caesarean section or known required referral, known severe congenital malformation (detected antenatally), intrauterine death or allergy to macrolides, and drugs known to prolong QT interval taken during the last 2 weeks, such as chloroquine, quinine, piperaquine, and erythromycin [18]. All women recruited in The Gambia who delivered a live-born neonate at a trial health facility were included in this secondary analysis.

Study procedures

Written informed consent to participate in the trial was taken during antenatal visits and verbally confirmed during labour. As per trial procedures, socio-demographic and perinatal clinical data were transcribed from the participants’ antenatal card and partograph (the timing of labour was not recorded in our study) to an electronic case report form (eCRF) before discharge by trained nurses [18]. This included the Apgar scores, and the status of the neonate at birth (alive or stillbirth) which were assigned contemporaneously by health facility nurses, midwives, or doctors. The delivery room health workers were trained on Apgar scoring relevance and methodology by research personnel shortly after the recruitment commenced, aiming to reduce potential subjectivity in scoring. Stillbirth was determined according to physical appearance, absence of heartbeat, and respiratory effort.

There was both active and passive surveillance by the trial team to collect information on the health status and evaluate the safety of the intervention for mothers and neonates [18]. A thorough clinical assessment of the mother and the neonates, including the neonate’s survival status at the time of assessment (alive or dead) was carried out by either a trained trial research clinician or a paediatrician 4–24 h after delivery [18] and before discharge. A further assessment was conducted at day 28 post-delivery, when mothers and their neonates were visited at home by trained trial nurses. Neonates requiring further examination by the trial clinicians were referred to the trial sites. There was passive follow-up of adverse events throughout the neonatal period, with mothers and neonates advised to visit the trial sites if they had health concerns. The trial clinician or a paediatrician then assessed the mother/newborn and decided to admit for further investigations (e.g. blood culture, full blood count) and management according to clinical judgment [18]. In the event of death, a serious adverse event report was written by the trial clinicians based on their assessment of the cause of death, using clinical notes and investigation results.

Outcomes and variables of interest

The primary outcome of this secondary analysis was the prevalence of IRA defined as a 1-min Apgar score of < 7, based on the ICD-10 version 2019 definition [20]. We chose 1-min Apgar score instead of 5-min Apgar score in recognition that the 1-min score reflects the intrapartum fetal health [11, 12] and we lacked detailed data about neonatal resuscitation efforts, hence 5-min scores were less reliable. We excluded neonates with an Apgar score of 7 as this is considered clinically reassuring and inconsistent with IRA [11, 21]. We defined severe IRA as a 1-min Apgar score between 0 and 3, and mild/moderate IRA as a 1-min Apgar score between 4 and 6 [20]. To explore factors associated with IRA, we developed an original conceptual framework based on a literature review and consideration of clinically and biologically plausible pathways of IRA in low-resource settings (Fig. 1) [22]. We differentiated the variables associated with IRA into two groups according to directionality: 1) Potential risk factors that may have contributed to the development of IRA (e.g., maternal, or obstetric complications); and 2) Factors that may have occurred due to the suspicion or presence of IRA (e.g., delivery location). The conceptual framework (Fig. 1) guided our choice of variables for the univariate (unadjusted) regression analysis and informed the adjusted analysis (see details below in Statistical analysis).

Fig. 1.

Conceptual framework for factors associated with intrapartum-related asphyxia. (Adapted from “Perinatal risk factors for neonatal encephalopathy: an unmatched case–control study | ADC Foetal and Neonatal Edition). Underlined = data available in the PregnAnZI-2 cohort. GBV: Gender-Based Violence. FGM: Female Genital Mutilation. GDM: Gestational diabetes mellitus. TORCH: congenital Toxoplasmosis, Others (Syphilis, Hepatitis B), Rubella, Cytomegalovirus (CMV), and Herpes simplex; GBS: Group B Streptococcus infections. TB: Tuberculosis. UTIs: Urinary Tract Infections. ANC: Ante Natal Care. HCW: Health Care Workers. CTG: Cardiotocography. C-section: Caesarean section

Operational definition

“Women with low-risk deliveries” refers to pregnant women in low-resource health systems such as The Gambia, where antenatal detection of high-risk women/neonates is challenging; this population is representative of pregnant women thought to be at low risk at the time of enrolment. Our study was embedded into a clinical trial that excluded “known” high-risk pregnancies (acute/chronic condition, planned caesarean section, severe congenital malformation, foetal macrosomia, etc.); in practice, often high-risk pregnancies were undiagnosed and were therefore included in the trial due to the lack of rigorous antenatal diagnosis in the trial health facilities. Participants with a history of previous caesarean section or who had undergone emergency caesarean section after enrolment were not excluded from the trial; only elective caesarean section of the current pregnancy was an exclusion criterion. Also, known major congenital malformations were an exclusion criterion. However, due to limited routine antenatal detailed ultrasonography at the study health facilities, congenital malformations were frequently not antenatally diagnosed and, therefore, were included in the trial.

Data management and statistical analysis

Relevant data was extracted from the PregnAnZI-2 trial electronic database and all analyses were done using R (version 4.1.2). Descriptive analyses were performed to determine prevalence estimates for all independent variables and adverse outcomes for the total cohort with stratification according to IRA status (low versus normal 1-min Apgar score). Considering the complexity and multifactorial nature of IRA, with diverse factors involved at different time points (pre-conception, antepartum, and intrapartum periods), an adjusted logistic regression model was developed for each independent variable found to have p-value < 0.2 for any of the categorical values on unadjusted regression. To avoid over-adjusting for multiple variables on the causal pathway [23], each model only included variables occurring before or at the same timepoint as the independent variable of interest (Fig. 1). We included events in previous pregnancies (i.e. stillbirths and miscarriages) in the risk factors analysis with the assumption that the underlying causes (e.g., infectious, genetic, maternal pre-existing morbidity) of these outcomes may be recurrent or persist during the current pregnancy [24, 25] with a potential risk for IRA.

Statistical significance for the adjusted analysis was pre-determined at p-value < 0.05.

Less than 1% of individuals were missing data in any variable and an available case analysis was done, i.e., individuals missing exposure or confounder variables were excluded from the adjusted analysis.

Results

Among 6758 live births included in the analysis (Fig. 2), 3.8% (257/6758) experienced IRA, with 79.0% (203/257) being mild to moderate IRA and 21.0% (54/257) severe IRA (Fig. 2). Details of baseline characteristics of the study neonates and their mothers are shown in Table 1. Overall, 11.7% (30/257) and 4.7% (11/257) of newborns with IRA had low birth weight and macrosomia, respectively; and 3.1% (8/257) of newborns diagnosed with IRA had a congenital malformation.

Fig. 2.

Flow chart for participants' selection process. m, mothers n, neonates

Table 1.

Maternal and neonatal risk factors for intrapartum-related asphyxia among low-risk pregnancies in peri-urban Gambia from October 2017 to May 2021

	Total (n = 6758)	Unadjusted univariate analysis				Adjusted analysis
		Yes (n = 257), n (%)	No (n = 6501), n (%)	cORa (95%CI)	p-value	aORb (95% CI)	p-value
Maternal factors
Age (Years)					0.42		0.99
Median (IQR)c	27 (23–31)
20–35	5459 (80.8)	203 (3.7)	5256 (96.3)	1		1
16–19	631 (9.3)	30 (4.8)	601 (95.2)	1.29 (0.86–1.88)		1.01 (0.67–1.58)
> 35	668 (9.9)	24 (3.6)	644 (96.4)	0.96 (0.61–1.45)		0.98 (0.63–1.61)
Mother’s Ethnicity					0.41
Mandinka	2701 (40.0)	115 (4.3)	2586 (95.7)	1
Fula	1213 (18.0)	36 (3.0)	1177 (97.0)	0.69 (0.46–0.99)
Wolof	1040 (15.4)	39 (3.8)	1001 (96.2)	0.88 (0.60–1.26)
Jola	876 (12.9)	31 (3.5)	845 (96.5)	0.82 (0.54–1.22)
Others	928 (13.7)	36 (3.9)	892 (96.1)	0.90 (0.61–1.31)
Father’s Ethnicity					0.29
Mandinka	2612 (38.7)	112 (4.3)	2500 (95.7)	1
Fula	1159 (17.2)	34 (2.9)	1125 (97.1)	0.67 (0.45–0.99)
Wolof	1165 (17.2)	40 (3.4)	1125 (96.6)	0.79 (0.54–1.14)
Jola	834 (12.3)	35 (4.2)	799 (95.8)	0.98 (0.65–1.43)
Others	987 (14.6)	36 (3.6)	951 (96.4)	0.84 (0.57–1.23)
Parity					< 0.001		< 0.001
Para 2	1392 (20.6)	50 (3.6)	1342 (96.4)	1		1
Para > = 4	2519 (37.3)	88 (3.5)	2431 (96.5)	0.97 (0.68–1.39)		0.80 (0.54–1.18)
Primiparous	1681 (24.9)	91 (5.4)	1590 (94.6)	1.54 (1.08–2.20)		1.69 (1.16–2.46)
Para 3	1166 (17.2)	28 (2.4)	1138 (97.6)	0.66 (0.41–1.05)		0.64 (0.39–1.03)
Season of deliveryd					0.50
Dry	4575 (67.7)	169 (3.7)	4406 (96.3)	1
Rainy	2183 (32.3)	88 (4.0)	2095 (96.0)	1.09 (0.83–1.43)
Previous stillbirthse					< 0.001		0.003
No	6642 (98.3)	245 (3.7)	6397 (96.3)	1		1
Yes	115 (1.7)	12 (10.4)	103 (89.6)	3.04 (1.50–5.64)		2.79 (1.41–5.52)
Previous caesarean section					0.06		0.03
No	6664 (98.6)	250 (3.8)	6414 (96.2)	1		1
Yes	94 (1.4)	7 (7.4)	87 (92.6)	2.06 (0.80–4.50)		2.48 (1.12–5.53)
Previous miscarriagef					0.09		0.02
No	6143 (90.9)	226 (3.7)	5917 (96.3)	1		1
Yes	613 (9.1)	31 (5.1)	582 (94.9)	1.39 (0.92–2.06)		1.65 (1.08–2.52)
Previous intake of traditional medicineg					0.30
No	5577 (82.6)	206 (3.7)	5371 (96.3)	1
Yes	1177 (17.4)	51 (4.3)	1126 (95.7)	1.18 (0.85–1.62)
Pre-rupture of membranesh					0.20
No	6600 (97.7)	248 (3.8)	6352 (96.27)	1
Yes	157 (2.3)	9 (5.7)	148 (94.3)	1.56 (0.69–3.09)
Neonatal factors	(n = 6758)				0.001
Sex							0.002
Female	3262 (48.3)	99 (3.0)	3163 (97.0)	1		1
Male	3496 (51.7)	158 (4.5)	3338 (95.5)	1.51 (1.16–1.97)		1.51 (1.16–1.97)
Birth weight (kg)i					< 0.001		< 0.001
Median (IQR)	3.1 (2.8–3.4)
Normal (2.5–4.0)	6085 (90.0)	215 (3.5)	5870 (96.6)	1		1
Low (< 2.5)	541 (8.0)	30 (5.5)	511 (94.5)	1.60 (1.06–2.34)		1.36 (0.88–2.08)
Macrosomia (> 4.0)	132 (2.0)	12 (9.1)	120 (90.9)	2.73 (1.41–4.82)		3.31 (1.78–6.15)
Plurality of birth					0.09		0.06
Singleton	6580 (97.4)	246 (3.7)	6334 (96.3)	1		1
Twin	178 (2.6)	11 (6.2)	167 (93.8)	1.70 (0.82–3.17)		1.90 (0.98–3.70)
Congenital malformationk					< 0.0001		< 0.001
No	6716 (99.4)	229 (3.5)	6366 (96.5)	1		1
Yes	42 (0.6)	14 (9.9)	128 (90.1)	3.04 (1.59–5.39)		3.01 (1.69–5.35)

^aCrude odds ratio,

^bAdjusted odds ratio,

^cInterquartile range,

^dDry season between November and May

^ePrevious stillbirth missing in n = 1

^fPrevious miscarriage missing in n = 2

^gPrevious intake of traditional medicine is any traditional medicine taken during the pregnancy irrespective of the route (oral, topical, or other), information on the exact period of use was not recorded, nor recorded what medicine, missing in n = 4

^hPre-rupture of membranes missing in n = 1

ⁱBirth weight missing in n = 2

^kCongenital malformations (minor and severe) diagnosed by the trial clinicians at birth—they were missed during the antenatal screening due to the limited access to antenatal diagnostic tools

a) Model 1 included the age of the mother, father’s ethnicity, mother’s ethnicity, parity, previous stillbirth, previous miscarriage, and previous caesarean section. b) Model 2 included the sex of the neonate, birth weight, plurality of birth, and congenital malformation

Factors associated with IRA

In the adjusted regression analysis, maternal risk factors for IRA included history of previous stillbirth (aOR: 2.79, 95%CI, 1.41–5.52, p-value = 0.003), previous caesarean section (aOR: 2.48, 95%CI, 1.12–5.53, p-value = 0.03) primiparity (aOR: 1.69, 95%CI, 1.16–2.46, p-value < 0.001), and previous miscarriage (aOR: 1.65, 95%CI, 1.08–2.52, p-value = 0.02) (Table 1). Neonatal risk factors were macrosomia (aOR: 3.31, 95%CI, 1.78–6.15, p-value < 0.001), congenital malformations (aOR: 3.01, 95% CI, 1.69–5.35, p-value < 0.001), and male sex (aOR: 1.51, 95% CI, 1.16–1.97, p-value = 0.002) (Table 1).

Several factors that may have occurred due to clinical suspicion of prolonged/difficult delivery or foetal distress were associated with IRA. Those with the highest ORs were delivery in referral hospitals (aOR: 17.9, 95%CI, 9.36–34.10, p-value < 0.001), deliveries attended by doctors (aOR: 3.25, 95%CI, 1.67–6.04, p-value < 0.001) and instrumental deliveries using forceps or vacuum (aOR: 2.92, 95%CI, 1.49–5.40, p-value < 0.001) (Table 2).

Table 2.

Association between intrapartum-related asphyxia and circumstances of delivery among low-risk pregnancies in peri-urban Gambia from October 2017 to May 2021

		IRA		Unadjusted		Adjusted
	Total	Yes (n = 257), n (%)	No (n = 6501), n (%)	cORa (95%CI)	p-value	aORb (95%CI)	p-value
Mode of delivery					< 0.001		0.29
Vaginal	6570 (97.2)	234 (3.6)	6336 (96.4)	1		1
Caesarean section	187 (2.8)	23 (12.3)	164 (87.7)	3.80 (2.30–6.03)		1.54 (0.70–3.34)
Assisted delivery					< 0.001		< 0.001
No	6255 (92.6)	212 (3.4)	6043 (96.6)	1		1
Episiotomy	426 (6.3)	28 (6.6)	398 (93.4)	2.01 (1.31–2.96)		1.65 (1.09–2.52)
Forceps/vacuum	77 (1.1)	17 (22.1)	60 (77.9)	8.08 (4.50–13.8)		2.95 (1.56–5.59)
Place of delivery					< 0.001		< 0.001
Health Centred	2864 (42.4)	27 (0.9)	2837 (99.1)	1		1
General Hospitalse	3782 (56.0)	205 (5.4)	3577 (94.6)	6.02 (4.09–9.22)		5.52 (3.66–8.32)
Referral hospitalsf	112 (1.6)	25 (23.17)	83 (76.9)	30.19 (116.8–54.30)		17.26 (9.10–32.74)
Staff conducting delivery					< 0.001		< 0.001
Midwife	5809 (85.9)	191 (3.3)	5618 (96.7)	1		1
Doctor	243 (3.6)	38 (15.6)	205 (84.4	5.45 (3.70–7.85)		3.26 (1.72–6.19)
Other	706 (10.5)	28 (4.0)	678 (96.0)	1.21 (0.79–1.79)		1.20 (0.81–1.85)

^aCrude Odds ratio

^bAdjusted Odds ratio

^cMode of delivery missing in n = 1

^dSerekunda Health Centre (SHC)

^eBundung Maternal and Child Health Hospital (BMCHH)

^fKanifing General Hospital (KGH) and Edward Francis Small Teaching Hospital (EFSTH)

All variables in the univariate analysis had a p-value < 0.2 and were all included in the adjusted analysis

Morbidity and mortality among newborns with IRA

Neonates with IRA were more likely to be hospitalised (OR: 11.91, 95% CI, 8.90–15.90, p-value < 0.001) and die (OR: 19.56, 95% CI, 12.02–31.55, p-value < 0.001) during the neonatal period than those without IRA (Table 3). Neonates with IRA represented 24% of all neonatal hospitalisations and 40% of all neonatal deaths (Fig. 3). This increased risk for death and hospitalisation was higher in severe IRA but still quite significant among mild/moderate IRA (Table 4).

Table 3.

Association between neonatal admissions/deaths and intrapartum-related asphyxia among low-risk pregnancies in peri-urban Gambia from October 2017 to May 2021

	Total	IRA (n = 257), n (%)	No IRA (n = 6501), n (%)	OR (95%CI)	p-value
Admission during the neonatal period					< 0.001
No	6370 (94.3)	164 (2.6)	6206 (97.4)	1
Yes	388 (5.7)	93 (24.0)	295 (76.0)	11.91 (8.90–15.90)
Neonatal mortality within 24 h after delivery					< 0.001
Alive	6733 (99.6)	237 (3.5)	6496 (96.5)	1
Died	25 (0.4)	20 (80.0)	5 (20.0)	109.25 (34.34–374.00)
Neonatal mortality between 24 h and 28 days					< 0.001
No	6413 (99.1)	211 (3.3)	6202 (96.7)	1
Yes	60 (0.9)	14 (23.3)	46 (76.7)	8.93 (4.47–16.87)
Overall mortality by day 28a					< 0.001
No	6413 (98.7)	211 (3.3)	6202 (96.7)	1
Yes	85 (1.3)	34 (40.0)	51 (60.0)	19.56 (12.02–31.55)

^aSurvival status by day 28 missing in n = 12

Fig. 3.

Prevalence of neonates with intrapartum-related asphyxia (IRA) at birth versus prevalence of neonates with IRA among neonatal admissions and deaths. A Prevalence of neonates with IRA at birth. B Prevalence of neonates with IRA among admissions (0–28 days). C Prevalence of neonates with IRA among neonatal deaths (0-24H after birth). D Prevalence of neonates with IRA among overall neonatal deaths (0–18 days)

Table 4.

Neonatal Hospitalisations and deaths associated with intrapartum-related asphyxia stratified by severity in a cohort of Gambian neonates born following low-risk pregnancies

	Severe (n = 54) n(%)	Normal n(%)	OR (95%CI)	p-value	Mild and Moderate (n = 203) n(%)	Normal n(%)	OR (95%CI)	p-value
Hospitalisation				< 0.001				< 0.001
No	28 (0.4)	6206 (99.6)			136 (2.1)	6206 (97.9)
Yes	26 (8.1)	295 (91.9)	19.5 (10.80–35.00)		67 (18.5)	295 (81.5)	10.35 (7.43–14.30)
Mortality within 24 h				< 0.001				< 0.001
No	40 (0.6)	6496 (99.4)			197 (2.9)	6496 (97.1)
Yes	14 (73.7)	5 (26.3)	448.5 (144.3–1600.00)		6 (54.5)	5 (45.5)	39.5 9.90–164.70)
Mortality between 24 h and 28 days				< 0.001				< 0.001
No	32 (0.5)	6202 (99.5)			179 (2.8)	6202 (97.2)
Yes	5 (9.8)	46 (90.2)	21.0 (6.10–57.80)		9 (16.4)	46 (83.6)	6.8 (2.9–14.3)
Overall mortality by day 28a				< 0.001				< 0.001
No	32 (0.5)	6202 (99.5)			179 (2.8)	6202 (97.2)
Yes	19 (27.1)	51 (72.9)	71.8 (36.00–140.40)		15 (22.7)	51 (77.3)	10.2 (5.20–18.80)

^asurvival status missing in n = 3 and n = 9 in severe and mild/moderate intrapartum-related asphyxia respectively

Discussion

IRA is a major cause of neonatal morbidity and mortality among neonates born in peri-urban health facilities in The Gambia, even for low-risk pregnancies. Although less than 4% (1 in 25) of neonates experienced IRA, these neonates represented 24% of all neonatal hospitalisations and 40% of all neonatal deaths in our cohort. We identified an important modifiable neonatal risk factor for IRA (i.e., macrosomia). In addition, we detected several maternal risk factors such as primiparity, history of previous stillbirth, or previous miscarriage that could signal a high-risk phenotype and trigger enhanced antenatal surveillance and early intervention to prevent IRA.

The prevalence of IRA in our cohort was lower than reported in other African studies, such as in Ethiopia (19.3%) [7] and Nigeria (29.4%) [26]. However, those studies were conducted at referral hospitals where high-risk deliveries, including caesarean sections, were more common, possibly explaining the higher prevalence of IRA. We also excluded women with known acute or chronic conditions [27] and, by selecting out women with detectable risk of intrapartum complications, this may have under-estimated the prevalence of IRA.

In our study, IRA was recorded in 2 out of 5 neonatal deaths, and 4 out of 5 neonatal deaths within 24 h of delivery. Therefore, the contribution of IRA to neonatal mortality is at least two- to three-fold higher than shown by previous studies in West Africa (such as Nigeria, 15.7%, and Ghana, 21.8%), despite the prevalence of IRA among neonates being higher [28, 29]. A possible explanation for the higher mortality in our setting compared to others in the sub-region is the lack of specialised care at delivery, both in terms of infrastructure and human resources. Indeed, in Nigeria, paediatricians and obstetricians were 24 h available for all high-risk deliveries [28]. In Ghana, paediatricians were present at all deliveries in which IRA was suspected, and all neonates at high-risk of neonatal encephalopathy were transferred to the Neonatal Intensive Care Unit, regardless of their Apgar scores [29]. This contrasts with our setting where trained midwives and nurses in the delivery room, medical officers were always available, but specialists (paediatricians and obstetricians) were inconsistently present due to their low number. As a result, delays in appropriate management of maternal and neonatal complications were occasionally observed. The risk for poor outcomes observed among neonates with IRA in our study was higher within the severe cases as expected but still significant among neonates with mild/moderate cases suggesting a possible subjectivity of the Apgar scoring and probable misclassification. Previous studies showed the same results [30]. This highlights that prompt and effective resuscitation and post-resuscitation management are critical to reducing mortality among neonates with IRA even in mild/moderate cases [31].

We identified both maternal and neonatal factors associated with IRA. From the maternal perspective, primiparity was an important risk factor for IRA. In line with previous studies in Nigeria [28] and Uganda [32], primiparous women often have prolonged labour, and are likely to have cephalo-pelvic disproportion, increasing the risk of obstructed labour [33]. Mothers with a history of previous stillbirth or miscarriage were also at higher risk of delivering a neonate with IRA. This clustering effect may be explained by either maternal biomedical factors (such as cephalo-pelvic disproportion), or biological factors (such as maternal chronic or acute conditions including infection, eclampsia, diabetes) which have been previously described in low- and middle-income countries [34–36]. From the neonatal perspective, male gender was associated with IRA; as previously described in Togo [37] and Ethiopia [38]. Although the mechanisms underlying this sex difference are not fully understood, females seem to have better in-utero hypoxic tolerance compared to male neonates [39–41]. Macrosomic neonates were also at higher risk of IRA, as previously described in studies from the region [42, 43]. Macrosomic neonates often present with peripartum complications such as obstructed second stage of labour [44, 45] or shoulder dystocia, highlighting the importance of foetal growth monitoring and planned Caesarean-section to mitigate this risk. Congenital malformation is a risk factor for IRA, as shown also in Tanzania [46] and Nigeria [47]. Although the exact causal pathway is not fully understood, a study in Finland found an association between congenital malformations and breech delivery [48]. It is well known that foetal malposition, especially breech presentation is associated with an increased risk of hypoxic-anoxic events [49] which may progress into IRA [50]. Early detection of foetal malposition by routine detailed sonography would enable planning for safe delivery. [51, 52].

Neonates with IRA were more likely to have been delivered in general and referral teaching hospitals, and were more frequent among deliveries conducted by doctors, consistent with studies conducted in Uganda [53] and Ethiopia [54]. In addition, neonates with IRA were more frequent among instrumental delivery (forceps or vacuum), similar to an Ethiopian study [55]. Reverse causality is the most likely explanation for these results. Indeed, complicated deliveries with high risk of IRA are indications for instrumental deliveries and these women are referred to highly equipped health facilities with more qualified staff such as midwives and medical doctors. These findings indicate that health workers were able to detect deliveries affected by IRA and make appropriate referrals and management decisions. However, these actions did not translate into improved neonatal outcomes as shown by the high mortality of neonates with IRA, reflecting the importance of timely decision-making, with reasons for delayed referrals or obstetric interventions likely to be multifactorial [56]. More research is required to understand the barriers and enablers for timely and optimal management of high-risk deliveries in varying West African settings to prevent avoidable neonatal deaths.

A limitation of this study is the potential for large intra- and inter-observer variability in Apgar score assignation, despite targeted training of health workers. The observed association between IRA and mortality indicates that specificity is probably high despite the potential for subjectivity. Secondly, overall neonatal mortality in our cohort is probably an underestimation of neonatal mortality due to both the healthy effect of research participation [57] and the exclusion of women with acute or chronic co-morbidities. The healthy effect of trial participation should have impacted all endpoints; therefore, decreasing the overall mortality rate but not necessarily the distribution of causes or risk factors. In summary, our findings are generalisable to a population of low-risk pregnant women delivering in low-resource West African hospital settings where antenatal detection of high-risk women and neonates is challenging yet may not apply to other populations or settings with more comprehensive antenatal diagnostics.

Conclusions

Our study highlights the high prevalence of IRA among neonatal deaths, even in a low-risk West African population, and several risk factors for IRA. Attention to modifiable risk factor (i.e., macrosomia) is required to decrease the prevalence of IRA. Better control of maternal diabetes and identification of foetal macrosomia for safe delivery plan are needed to modify the association of macrosomia with IRA. In addition, effective interventions to improve antenatal and obstetric care and post-delivery management of neonates exposed to IRA, such as prompt resuscitation by trained, skilled staff, are crucial if global neonatal mortality reduction targets are to be met by 2030.

Abbreviations

IRA

Intrapartum-Related Asphyxia

DALY

Disability Adjusted Life Years

SSA

Sub-Saharan Africa

SDG

Sustainable Development Goal

SHC

Serekunda Health Centre

BMCHH

Bundung Maternal and Child Health Hospital

ecrf

Electronic Case Report Form

Authors’ contributions

AR and UdA conceived and designed the initial trial. HT contributed to the trial design and implementation. NB designed this study with input from AR, HB, and CB. NB performed the analysis with full access to the data and input from CB. NB drafted the initial manuscript with input from HB, AR, CB, FS, and UdA. NB, AR, HB, CB, HT, BC, UNN, FS, MD, JCJ, SG, EN, YN, ES, SB, OJ, AS and UdA contributed to the writing of the final version, and have read and approved it. AR gave oversight to the work as guarantor and accepts full responsibility for finished work and controlled the decision to publish. All co-authors gave their approval for the publication of the final version of the manuscript.

Funding

The PregnAnZI-2 trial was funded by a grant from the UKRI under the Joint Global Health Trial Scheme (JGHT) (ref: MC_EX_MR/P006949/1) and support from the Gates Foundation (Ref:OPP1196513). The publication of this article was funded in whole by the Gates Foundation. The funders and study sponsor (MRCG) had no role in the study design, collection, analysis, or interpretation of data, writing of article nor the decision to submit for publication.

Data availability

Data may be obtained from a third party and are not publicly available. The clinical data has been collected following provision of informed consent under the prerequisite of strict participant confidentiality. Qualified researchers may request access with the Gambia Government/MRC Joint Ethics Committee. The review process and release of data will be facilitated by MRC Unit The Gambia (http://www.mrc.gm/) through the Head of Governance at MRCG. Access will not be unduly restricted.

Declarations

Ethics approval and consent to participate

This study is a secondary analysis of the PregnAnZI-2 clinical trial. The trial was approved by The Gambia Government/Medical Research Council Unit The Gambia (MRCG) Joint Ethics Committee, the Comité d’Ethique pour la Recherche en Santé (CERS) and the Ministry of Health of Burkina Faso, and the London School of Hygiene and Tropical Medicine (LSHTM) Ethics Committee. This included approval for further research utilising trial data. During antenatal care visits all participants provided written informed consent to participate in the trial and were free to withdraw at any time. They also consented for the use of their data for further research.

Consent for publication

Not applicable.

Competing interests

Dr Usman Nakakana is current employee of Gates Foundation but he doesn't own shares and was not affiliated with them at the time the study was conducted.