Anterior abdominal wall defect is a spectrum of congenital malformations arising from defective fusion or development of the embryonic folds [1] . These include—omphalocele, gastroschisis, ectopia cordis, bladder exstrophy, exstrophy complex, limb-body wall complex, body stalk anomaly, amniotic band syndrome, and prune belly syndrome. These defects could present with the least morbidity, such as in a bifid clitoris, to the most severe morbidity, such as seen in cloacal exstrophy and pentalogy of Cantrell. The incidences of these malformations vary depending on the type of defect, with gastroschisis occurring in 1:2000 live births, followed by omphalocele with the incidence of 1:4000 live births [2] [3] .

Anterior abdominal wall defects present a conundrum of management decisions to the surgeon and with varying outcomes depending on so many factors. Studies have evaluated the management decisions that offer the best possible outcomes [4] . These outcome measures include length of hospital stay, duration of use of mechanical ventilators, and dependence on parenteral nutrition [5] . Survival is well above 90% in developed nations. In the developing world, however, with a severe lack of supportive equipment like mechanical ventilators and parenteral nutrition, outcomes are measured based on survival, which is abysmal, ranging from 50% to 0% [6] - [8] .

Studies have shown that survival in newborns with anterior abdominal wall defects is a function of many factors, which include the type of defects, associated chromosomal, cardiac defects, and sepsis [9] [10] . Mortality is higher in syndromic omphalocele major with associated chromosomal abnormalities, as against simple non-syndromic omphalocele [11] [12] . Complex gastroschisis has a relatively higher mortality than simple gastroschisis [12] [13] . Sepsis is another significant factor in mortality. However, in low- and middle-income countries (LMICs) like our setting, where facilities for chromosomal analysis, cardiac evaluation, nutritional support and mechanical ventilator are not readily available, mortality is still very high [7] [14] .

Patient, maternal, and management variables predicting mortality among neonates with anterior abdominal wall defects have not been fully investigated in our setting. This study aimed to look at these variables that could predict mortality in our environment. The findings could offer guidance to clinicians, health managers, and policy makers in designing measures geared toward reducing mortality in newborns with anterior abdominal wall defects.

A total of 142 newborns presented to our facility with anterior abdominal wall defects. Females were 74 and males 68, with a male-to-female ratio of 0.9:1. There were 113 (79.58%) booked pregnancies and 29 (20.42%) unbooked. The median age at presentation was 18.5 hours while the mean birth weight was 2.6 ± 0.6 kg. Median weight was 2.7 kg. Their ages ranged from 1.1 kg to 4.2 kg. There were 120 (83.8%) newborns with omphalocele, 14 (9.9%) with gastroschisis, prune belly 3 (2.1%), bladder exstrophy, cloacal exstrophy, a set of conjoined twins and a pentalogy of Cantrell 1 (0.7%) each. Of the 120 with omphalocele, 107 were omphalocele major, 13 omphalocele minor and 45(37.5%) presented with ruptured membrane. The complication recorded was only sepsis in 38,26.7%. The length of hospital stay ranged from 0.2 days to 63 days, with a mean of 15 days. Those who were discharged home had a mean length of hospital stay of 21.7 days while the mean length of stay for those who died was 6.1 days. There were 62, 43.7% mortalities in the study group while 80, 56.3% were treated and discharged home.

Out of the 62 mortalities, 45 (72.6%) occurred within the first week of admission. The mean length of hospital stays and birth weight in this group who died within the first week were 3.5 days and 2.3 kg, respectively.

There were 49 (34.5%) neonates with low birth weight. Out of this, 39 (79.6%) were products of booked pregnancy (at primary healthcare 11, 30%, private facility 5, 11% and secondary and tertiary healthcare centres 23, 59%). Among the low birth weight babies, the mean gestational age was 32 weeks, with a range of 28 - 36 weeks. Mortality among the booked low birth weight pregnancies was 29, 74.36% (among booked primary healthcare centres 11, 72.2%, private facilities 3, 60% and secondary and tertiary healthcare centres 18, 78.26%). Case fatality rates were: pentalogy of Cantrell 1 (100%), Gastroschisis 13 (92.85%) and omphalocele major 44 (41.1%) (see Figure 1 ).

Of the 44 mortalities with omphalocele major, 26 (51.1%) had ruptured membranes.

Using bivariate analysis, Birth weight (p < 0.001), Type of defect (p = 0.001), Maternal age (p = 0.002) and length of hospital stay (p < 0.001) were significantly associated with mortality (see Table 1 ).

These variables were further subjected to logistic regression which revealed that birth weight (OR = 0.209, CI = 0.083 - 0.525, p = 0.001); Duration of hospital stay (OR = 0.287, CI = 9.985, p < 0.001) and type of defect (OR = 0.449, CI = 0.276 - 0.731, p = 0.001) significantly predicted mortality. Although mortality increased with sepsis, the difference was not statistically significant (OR=1.765, CI=0.829 - 3.756, p = 0.141) ( Table 2 ).

In our study population, birth weight, type of defect, and duration of hospital stay were significant predictors of mortality. Maternal age was significantly associated with mortality, but was not a significant predictor of mortality.

Mortality was highest in newborns with low birth weight and gestational age of less than 37 weeks. This may be attributed to the various complications associated with low birth weight, which include hypothermia, sepsis, and the associated congenital anomaly. According to the World Health Organisation’s global report on preterm birth, 2012, low birth weight newborns have a higher risk of dying [15] . Another study in Nigeria also found that mortality was high in newborns with low birth weight [16] . Biswas I et al. in Bangladesh, however, had a different outcome, survival was higher in term and normal birth weight babies, although they noted it was not statistically significant [17] . However, a study carried out in a level III NICU in Japan found that mortality was just 11% among low birth weight babies [18] . This improved survival outcome compared to our study could be due to access to the high-level care given to these patients.

Antenatal care: The majority of the pregnancies were booked. Among the low birthweight and preterm babies, we found that more than 2/3^rd were booked pregnancies. However, getting booked did not translate to improved survival. Anterior abdominal wall defects can be diagnosed prenatally [19] [20] . Prenatal diagnosis can guide tailored management and improve survival. The reason for the high mortality rate among booked pregnancies in our study could be due to poor implementation of an antenatal protocol for early detection of congenital anomalies.

The Type of defect was also found to be a significant predictor of mortality in the study group. Mortality was mainly in newborns with ruptured omphalocele and gastroschisis. Exposure of the bowel while moving from one facility to another could have increased their risk of hypothermia, fluid and electrolyte loss, and sepsis. Bowel function could have also been compromised, while this could have been easily overcome with parenteral nutrition, it is not readily available. In a poorly equipped neonatal intensive care unit, the outcome was likely to be poor.

In our previous study the use of improvised preformed silo (femidom) had reduced mortality in ruptured omphalocele and gastroschisis from 100% to 50% [6] . Similar studies in Africa found mortality in newborns with gastroschisis to be abysmally high (70% - 100%). They found that, lack of prenatal diagnosis and an inadequately equipped neonatal intensive care increased the risk of dying [7] [14] . The use of improvised silo alone is not enough to reduce mortality. This is in contrast to developed nations, where mortality is as low as 0% - 2% [5] . Ruptured omphalocele complicates management, increases morbidity, and contributes to the significantly high mortality rate in these patients [21] .

Although mortality increased in newborns with sepsis, there was no significant correlation between sepsis and mortality in this study. Although other studies have found sepsis to be the primary cause of death in these patients [22] .

Length of hospital stay was a significant predictor of mortality. Mortality was highest within the first week of admission, and patients who lived beyond 2 weeks were most likely to survive. Although the majority of our patients presented to our facility relatively early within the first 10 hours after birth and received the same resuscitative measures, mortality remained high in the first week. This emphasizes the need for prompt intervention immediately after birth,

Age at presentation, sex, Maternal parity, and maternal age did not predict mortality in our study. A study by Schimmel et al. reported a similar finding, that maternal age and parity did not affect mortality, but it was associated with small for gestational age babies, and preterm delivery [23] . In their study, however, Kazuki et al. reported that maternal age and parity were associated with small for gestational age, preterm delivery, and mortality [24] . The difference in our study could be the variations in the ages and parity of mothers in the study groups. Similar age and parity may produce similar results.

Our finding of low birth weight and type of defect (ruptured omphalocele and gastroschisis) as significant predictors of mortality, emphasises the need to improve our antenatal care services for prenatal diagnosis and provision of neonatal intensive care could reduce mortality.

1) Being a retrospective study, some data on the extent, compliance, and skills of attending health personnel among booked pregnant mothers could not be accessed.

2) Resource limitation: Lack of adequately equipped intensive care unit and parenteral nutrition which could impact outcome.

3) In ability to evaluate for other confounders like chromosomal abnormalities and cardiac defects that could influence outcomes.

These limitations could have had an impact on the outcome of our study. A prospective study taking into consideration these limitations would have reduced bias.