Discussion

In the present study on neonatal mortality among LBW infants in two big Malawian hospitals, we found an overall neonatal mortality rate of 17%, reaching 48% and 93% in the groups of neonates with birth weights 1000–1499 g and <1000 g, respectively. For babies <1200 g only, neonatal survival was higher in the tertiary hospital with better personnel and equipment resources. Most deaths occurred within the first 3 days of life. The coverage of antenatal steroids was less than 10%. A small minority of newborns were treated with CPAP, in most cases without any positive effect.

Few studies have been made on the survival of LBW in similar settings in other low-income countries. Studies published in the 1990s showed similar rates of survival in Ghana and Nigeria with survival of ELBW of 10%–17%, of VLBW of 47% and LBW of 89%.10 ,11 Newer reports from less resource poor settings, but without access to surfactant and ventilators, show survival of up to 30% for ELBW, 67%–83% for VLBW and 98% for LBW.12 ,13 Hence there is room for improvement in KCH and BH.

Most babies below 1000 g need neonatal intensive care to survive.14 As there are no neonatal ventilators and no surfactant available in BH or KCH, the poor outcome of the ELBW babies was expected.

Even though the survival rate increased with increasing weight, in numbers, equally as many babies below as above 1500 g died in this study. A majority of deaths occurred within the first 3 days of life. The most common causes of death within the first few days of life in preterm infants are asphyxia, hypothermia, hypoglycaemia, neonatal sepsis and respiratory distress syndrome (RDS).15 The majority of these conditions can be treated or prevented with relatively simple means by resuscitation, early and continuous skin-to-skin contact to prevent hypothermia, improved early-feeding practices to prevent hypoglycaemia, antibiotics to treat infections and antenatal corticosteroids, oxygen and access to simple CPAP treatment to prevent and treat RDS.8 ,15–17 Improvement in these areas would likely improve the survival of the heavier babies significantly. As 40% of the deaths in BH happened within the first day (compared with 16% in KCH), special efforts should be put into improving resuscitation measures in BH. In both hospitals, the KMC and the early feeding are often delayed by a practise of wrapping the baby in cloths and putting it aside after successful resuscitation. Sometimes, this is due to complications of the mother that need to be dealt with and not enough staff. However, most mothers have an accompanying guardian who could give skin-to-skin contact already in labour ward and take the baby to the nursery for feeds and further care within the first hour.

As almost all CPAP treated babies in this study died, the use of the CPAP machine and the timing of initiating CPAP treatment need to be revised. As the machine is not used on a daily basis and there is a high turnover of staff, there might be improvement by offering training for the staff on a regular basis. Given the limited resources in BH and KCH and the high risk of sequelae for the smallest babies, this study suggests that efforts should currently be focused on the infants with a birth weight of more than 1100 g (figure 1).14 ,18

KCH had, as the tertiary hospital, a higher proportion of ELBW and VLBW, resulting in a higher overall mortality rate. Despite the fact that KCH received many of the smallest and sickest babies from BH and other facilities as referrals, the survival was equal or better than in BH for babies below 2000 g. Access to paediatricians, better equipment and more nursing staff per baby in KCH might have contributed to the better outcome. One could also argue that a baby that survived the transfer to KCH, which sometimes can take hours from some of the district hospitals, has somehow already proven to be strong.

There are currently no guidelines agreed upon for referrals of babies with LBW between the two hospitals in Lilongwe. If the tertiary level hospital had enough resources in terms of beds, staff and supplies, it would probably be wise to centralise the care of all infants below 1500 g. Ultimately, those babies should be transferred in utero. However, with the current limited resources and the very poor outcome for infants below 1100 g, we suggest referral of babies weighing 1100–1200 g to the tertiary level.

The CS rate of 18% is higher than the 5%–15% recommended by WHO and may have affected the survival of the babies (table 1).19 The national rate of CS is 4.6%.3 The high rate of CS in the central hospitals reflects the poor health status of these mothers and the high percentage of multiple gestations. In each case of provider initiated delivery and CS, the obstetrician must balance the mother's condition with the risks that prematurity mean to the baby and surgery to the mother.

Antenatal corticosteroids significantly reduce neonatal morbidity, mortality and need for respiratory support in preterm infants.20 Surprisingly, this study did not support any conclusions on the efficiency of antenatal corticosteroids. However, the numbers were very small, the average weight of the exposed babies was slightly lower than of the non-exposed and few of the exposed babies had a full course (and therefore the full benefit) of the antenatal corticosteroids. Many mothers were admitted when in active labour and delivered shortly after admission. As a tertiary centre, KCH had a higher proportion of mothers who were admitted due to complications of pregnancy, for example, preeclampsia. The longer admission times prior to delivery for the sick mothers in KCH may have allowed time for administration of antenatal corticosteroids, but their disease might also have added to the risk of morbidity and mortality of their babies.

In this study, less than 10% of the babies had a mother who had received antenatal corticosteroids and only 4% had had a full course. The coverage was lowered by the fact that during the study period, corticosteroids were intermittently unavailable in the pharmacies of the two hospitals. The availability and use of antenatal corticosteroids need to increase. Possibly the coverage would increase if injectable corticosteroids are stocked in health centres and district hospitals, where mothers at risk of preterm delivery could be treated before referral to the maternity units.

From an obstetric point of view, persistent malaria, maternal undernourishment and anaemia have been associated with preterm birth in Malawi.21 More research is needed to investigate the causes of premature birth and LBW and why Malawi has such high rates of especially preterm birth. Improved quality of the components of the antenatal clinics and focus on risk factors for preterm birth could possibly help to identify the mothers at risk and enable referral to the most adequate level of care for them.

We acknowledge some weaknesses of the present study. The study was conducted in the two hospitals in the capital city, which have a higher standard of care than most other hospitals in the country, and hence the findings cannot be generalised for all of Malawi. The preterm birth rate could not be calculated in this study, as both inborn and outborn babies were included. We only investigated the short-term survival of the infants with LBW. We have no information on the causes of death or the long-term survival or morbidity, which in many cases might be lifelong and be a heavy burden on the individuals, the families and health systems.18 ,22 These issues need to be addressed to reduce the under-five mortality in Malawi.