Is Late-onset Gram-negative Bloodstream Infection Associated with Gastrointestinal Pathology in Infants Hospitalized in Neonatal Units?

    loading  Checking for direct PDF access through Ovid


The Matching Michigan approach to nosocomial infection prevention in intensive care units has largely focused on improving central line care.1 However, late-onset infection in neonatal intensive care units may not always be catheter-associated. We explored whether there is an association between late-onset Gram-negative bloodstream infection (GNB BSI) and abdominal pathology.
We performed a retrospective audit of all infants with late-onset GNB BSI admitted to the John Radcliffe Hospital, Oxford, United Kingdom, from March 2011 to June 2016. Late-onset BSI was defined as occurring after 48 hours of life. Abdominal radiographs taken around the time of infection (5 days before, to 8 days after positive blood culture) were assessed for the presence or absence of gastrointestinal pathology.
Forty-two episodes of late-onset GNB BSIs occurred in 40 infants. In 35 infection episodes, an abdominal radiograph had been performed, and in 26 (62% of the total), there was evidence of gastrointestinal pathology.
In 16 infection episodes, either an abdominal radiograph was performed and thought to be normal (9 infections) or no abdominal radiograph was taken (7 infections). Hence, in 38% of GNB BSIs, there was no obvious abdominal source for the infection.
Seven patients (17%) had positive respiratory tract cultures for the same organism (mostly bronchoalveolar lavage). None of the 42 episodes were associated with a positive urine culture.
There was no convincing clinical history indicating that the central venous line (CVL), if present, was the cause of the infection. Of the 26 episodes where there was gastrointestinal pathology, 21 (81%) were associated with a central line in situ. In comparison, 9/16 (56%) infection episodes not associated with abdominal pathology had central lines in situ at the time of the infection. CVLs appeared to be casually rather than causally correlated with the infection episode, and a CVL was more often present when the infant had evidence of gut pathology.
Based on these findings, we suggest an association between late-onset GNB BSI and gastrointestinal pathology, and where there is no gut pathology, the source of the infection is likely to be ventilator-associated pneumonia (VAP). No GNB BSI occurred secondary to a urinary tract or CVL infection. Therefore, reducing necrotizing enterocolitis and VAP appears to be a key to preventing these infections. The Matching Michigan infection reduction program has been successful in reducing intensive care unit nosocomial infections, but the original emphasis on improved central line insertion and post-insertion hub care may not have directly transferrable relevance to patients in neonatal units. We need to consider other means of infection prevention, aimed at reducing necrotizing enterocolitis and VAP, if we are to effectively reduce GNB infection in infants.

Related Topics

    loading  Loading Related Articles