We conducted a retrospective analytical cohort study in the neonatal unit of Yalgado Ouédraogo University Hospital, the main national tertiary referral center in Burkina Faso from October 1st, 2021 to September 30th, 2024.

All newborns admitted to the neonatal unit during the study period with a birth weight <1000 g were eligible. They were either born in the maternity ward of the hospital (inborn) or referred from another health facilities (outborn).

Birth weight <1000 g.Admission to the neonatal unit.Available and complete medical record.

Missing key clinical data.Prior hospitalization elsewhere.

The variables studied included maternal variables (age, number of pregnancies, prenatal consultations, residence, socioeconomic status) and neonatal variables (weight, gestational age, sex, respiratory distress, temperature). Gestational age was estimated using the last menstrual period, obstetric ultrasound when available, or Valery Farr score. Respiratory distress was defined as the presence of at least two of the following criteria: change in respiratory rate (tachypnea >60 breaths/min or bradypnoea (<30 breaths/min); signs of pulmonary retraction (nasal flaring, intercostal retraction, xiphoid retraction, expiratory grunting, poor thoracoabdominal synchronization); cyanosis; oxygen requirement.

Hypothermia was defined as axillary temperature <36.5˚C and fever has axillary temperature ≥37.5˚C according to WHO criteria.

The main outcome variable was hospital survival.

Data were analyzed using STATA version 15. Missing data were assessed for each variable. Variables with more than 10% missing data were excluded from multivariable analysis. Cases with missing values for selected predictors were excluded using a complete-case analysis approach. Continuous variables were expressed as mean ± SD, and categorical variables as frequencies and percentages. Survival probability was estimated using the Kaplan-Meier method with birth as the starting time, death as the event, and discharge alive as censoring. Although time-to-event data were available, logistic regression was used to identify independent predictors of in-hospital mortality because the primary objective was to evaluate overall mortality during hospitalization rather than time-dependent hazards. A Cox proportional hazards model produced similar results (data not shown).

Factors associated with mortality were identified using multivariate logistic regression. Statistical significance was set at p < 0.05. Variables with p < 0.20 in univariate analysis and clinically relevant variables were included in the multivariable logistic regression model. Collinearity between predictors was assessed using variance inflation factors.

During the study period, 110 extremely low birth weight newborns were admitted among 1528 neonatal hospitalizations,representing 7.2% of all neonatal admissions.

• Mothers’ characteristics

Most mothers were aged between 18 and 35 years (82.7%). Rural residence was reported in 67.3% of cases and 87.3% had a low socioeconomic status. More than half of mothers (57.3%) had attended only 2 - 3 antenatal care visits, and only 6.3% had ≥4 visits (Table 1).

Table 1. Distribution of newborns according to the sociodemographic and obstetric characteristics of their mothers.

The mean birth weight was 810 g (500 - 990 g) and mean gestational age was 29 weeks (22 - 34 weeks). Female newborns accounted for 61% of cases. The median time to admission was 4.4 hours, ranging from 30 minutes to 40 hours. Clinical conditions included respiratory distress: 68.2% and hypothermia: 45.5% (Table 2).

Table 2. Newborns characteristics (n = 110).

Overall hospital mortality was 71% (78/110).

Kaplan–Meier analysis showed a rapid decline in survival during the first days of life with an estimated probability of 72% at 24 hours, 48% at 7 days, and 29% at 28 days. The median survival time was 4 days, indicating a very high rate of early neonatal mortality. More than 90% of deaths occurred during the first week of life (Figure 1).

Male sex (OR = 3.92) and respiratory distress (OR = 3.39) increase the risk of death, while weight ≥ 900 g (OR = 0.2) is a protecting factor for survival (Table 3).

Figure 1.Kaplan-Meier survival curve of extremely low birth weight newborns (<1000 g).

Table 3. Factors associated with mortality.

The retrospective and single-center nature of the study, as well as the lack of follow-up after discharge, are the main limitations of the study. Despite these limitations, it enabled the assessment of hospital survival and the identification of determinants of mortality among extremely low birth weight newborns hospitalized in a tertiary hospital in a resource-limited country.

This study highlights the extremely high mortality of ELBW newborns in a tertiary hospital in sub-Saharan Africa. All newborns in our cohort were extremely preterm. The mean gestational age was 29 weeks, indicating extreme prematurity. The observed mortality rate (71%) is consistent with previous African studies reporting mortality rates exceeding 60% - 80% among infants weighing less than 1000 g [8]-[12].

In contrast, survival rates in high-income countries exceed 80% due to advances in neonatal intensive care, including mechanical ventilation, surfactant therapy, and specialized nutritional support [5][6].

The Kaplan–Meier survival analysis demonstrated that mortality occurred predominantly during the early neonatal period, with more than 90% of deaths occurring within the first week of life. This pattern has been widely reported in low-resource settings and reflects the vulnerability of ELBW newborns to respiratory distress, hypothermia, hypoglycemia sepsis, and feeding difficulties. Indeed, neonatal infections also contribute significantly to mortality among very low birth weight infants in resource-limited settings. Furthermore, nutritional management is another key determinant of survival among very low birth weight infants [7]-[10].

Respiratory distress emerged as a major determinant of mortality. This finding is consistent with previous studies indicating that respiratory distress syndrome remains one of the leading causes of death among extremely premature infants [10].

The absence of advanced respiratory support modalities such as CPAP and surfactant therapy likely contributed significantly to the high mortality observed in our setting. Evidence from Malawi and other African countries has shown that the introduction of low-cost CPAP systems can substantially improve survival among preterm infants [7]-[9][11][12][15].

Birth weight was another strong determinant of survival. Infants weighing ≥900 g had significantly better outcomes than those with lower birth weight. This relationship between birth weight and neonatal survival has been consistently demonstrated in the literature [3][6][12][13].

Male sex was associated with increased mortality, confirming the well-documented biological vulnerability of male newborns during the neonatal period [16].

Overall, these findings reflect the combined impact of limited neonatal intensive care resources, insufficient antenatal corticosteroid use, frequent postnatal transfers, and shortages of trained healthcare personnel. Strengthening the continuum of care from pregnancy to neonatal intensive care is therefore essential to improve survival outcomes in this population [17][18].