Predictors of mortality, length of stay and co-morbid hypothermia in hospitalized neonates with pneumonia in Eritrea, Africa.

Pneumonia is the third leading cause of neonatal death in Sub-Saharan Africa, however systematic reviews of newborn mortality in this region are lacking [1–6]. The majority of cases occur in developing world countries, which have no formal reporting system for these deaths, many of which occur at home [7, 8]. We sought to determine the predictors of mortality in a hospitalized cohort of infants with pneumonia in Asmara, Eritrea, conducting a retrospective review of all 2006 admissions to the Orotta Pediatric Hospital Neonatal Intensive Care Unit, the nation’s only tertiary newborn centre. Independent for less than three decades, the country introduced a ‘Specialized Neonatal Care Unit’ (SNCU) in 2003. The SNCU was attended to by two physicians and provided a centralized oxygen supply, infusion pumps and incubators, three radiant warmers and phototherapy equipment. Bubble Continuous Positive Airway Pressure was the primary means of assisted ventilation [9]. Data on age, birth weight, gender, mode of delivery, Apgar score, maternal age, birth location, admission diagnosis, admission comorbidities and outcome was collected. Multivariate analysis determined predictors of mortality and length of stay. Bivariate analysis determined the impact of admission hypothermia (temperature <36.5 C) on outcomes [10]. Neonatal pneumonia was a clinical diagnosis informed by the presence of radiographic infiltrate(s) or focally decreased breath sounds on examination with clinical evidence of increased work of breathing, laboratory leukocytosis and/or bandemia [11]. Of 1503 infants analyzed, neonates with pneumonia had younger gestational ages, lower birthweights, lower Apgar Scores, presented to care earlier in life, had longer lengths of stay and mothers of younger age than those without. Mortality from neonatal pneumonia was 13.7%, (Fig. 1). Gestational age, Apgar scores (p< 0.001 for each), and, birthweight <2 kgs, (p< 0.04) predicted mortality (Table 1 Regression). Hypothermia was defined as a temperature of <36.5 C rectally at the time of first medical presentation, including rural medical facilities prior to admission. Hypothermia complicated 25.8% (79/306) of cases with pneumonia (Fig. 2). Hypothermia was inversely related to birthweight, occurring in 50% of patients 2 kgs. For infants with a birthweight of 2 kgs with pneumonia, mortality was higher in infants with co-morbid hypothermia (p< 0.05, Fig. 3). No mortality was seen in hypothermia patients >2.5 kgs. No significant differences in mortality were seen in normothermic infants across weight classes >2.0 kgs. Multivariate regression modeling was performed to determine association with length of stay. Decreased gestational age and birthweight 2 kgs were associated with increased length of hospitalization (p< 0.05 and p< 0.0001, respectively, Table 2). This work is the first major report on neonatal pneumonia from this region. Focus on thermal care may reduce mortality in neonates with pneumonia< 2 kgs. Efforts to prevent hypothermia can focus on support for skin-to-skin contact and maintenance of the ‘warm chain’ [12–14]. Given the discrepancy in mortality rates from pneumonia between low birthweight, hypothermic and normothermic newborns, modest improvements could result in substantial mortality reductions. Improvements