Practice variations and rates of late onset sepsis and necrotizing enterocolitis in very preterm born infants, a review

Mark Adams, Dirk Bassler


The burden of late onset sepsis (LOS) and necrotizing enterocolitis (NEC) remains high for newborns in low- and high-income countries. Very preterm born infants born below 32 weeks gestation are at highest risk because their immune system is not yet adapted to ex-utero life, providing intensive care frequently compromises their skin or mucosa and they require a long duration of hospital stay. An epidemiological overview is difficult to provide because there is no mutually accepted definition available for either LOS or NEC. LOS incidence proportions are generally reported based on identified blood culture pathogens. However, discordance in minimum day of onset and whether coagulase negative staphylococci or fungi should be included into the reported proportions lead to variation in reported incidences. Complicating the comparison are the absence of biomarkers, ancillary lab tests or prediction models with sufficiently high positive and/or negative predictive values. The only high negative predictive values result from negative blood culture results with negative lab results allowing to discontinue antibiotic treatment. Similar difficulties exist in reporting and diagnosing NEC. Although most publications base their proportions on a modified version of Bell’s stage 2 or 3, comparisons are made difficult by the multifactorial nature of the disease reflecting several pathways to intestinal necrosis, the absence of a reliable biomarker and the unclear differentiation from spontaneous intestinal perforations. Comparable reports in very low birthweight infants range between 5% and 30% for LOS and 1.6% to 7.1% for NEC. Evidence based guidelines to support treatment are missing. Treatment for LOS remains largely empirical and focused mainly on antibiotics. In the absence of a clear diagnosis, even unspecific early warning signals need to be met with antibiotic treatment. Cessation after negative blood culture is difficult unless the child was asymptomatic from the beginning. As a result, antibiotics are the most commonly prescribed medications, but unnecessary exposure may result in increased risk for mortality, NEC, further infections and childhood obesity or asthma. Finding ways to limit antibiotic use are thus important and have shown a large potential for improvement of care and limitation of cost. Over recent decades, none of the attempts to establish novel therapies have succeeded. LOS and NEC proportions remained mostly stable. During the past 10 years however, publications emerged reporting a reduction, sometimes by almost 50%. Most concern units participating in a surveillance system using quality improvement strategies to prevent LOS or NEC (e.g., hand hygiene, evidence based “bundles”, feeding onset, providing own mother’s milk). We conclude that these approaches display a potential for wider spread reduction of LOS and NEC and for a subsequently more successful development of novel therapies as these often address the same pathways as the prevention strategies.