Setting and Population
The “Civico” hospital NICU is set in the adult hospital campus and includes two adjacent open spaces of 60 and 40 square meters respectively. The NICU has 16 beds, of which 8 intensive and 8 sub-intensive. A hand-washing sink is located at each entrance. Around 370 term and preterm newborns are admitted annually.
Each area, intensive and sub-intensive, has two dedicated neonatologists and two dedicated nurses, while another nurse is responsible for feeding-preparation.
All newborns included in our monthly surveillance program between November 2016 and October 2019 were enrolled in this study. Rectal swabs for detection of MDR-GNB carriage were collected every 4 weeks from all newborns in the NICU, regardless of any clinical or laboratory signs of infection. Carriage was defined as a positive culture of MDR-GNB from at least one rectal swab collected during the NICU stay.
Since November 2017, a strategy of multiple coordinated intervention measures was introduced to reduce the prevalence of MDR-GNB carriage.
Description of Intervention Measures
The intervention strategy included:
a) Intensification of sample collection
Microbiological surveillance was reinforced for 2 months (11/27/2017 - 01/23/2018) as follows:
- additional weekly samplings of rectal, nasal and oral mucosa swabs, swabs of devices and material in direct contact with each newborn (feeding bottles and pacifiers, and remnant milk after newborn feeding) for the first month and every four weeks afterwards;
- environmental samplings (room surfaces including milk preparation room, healthcare workers hands and stethoscopes) at the beginning of the intervention and after two months. In February 2018 monthly sample collection reverted to the previous surveillance with only rectal swabs.
b) Stakeholders weekly meetings
During the first two months of intervention, weekly meetings of NICU healthcare workers were held with experts from the surveillance team, sharing surveillance program results, pinpointing adherence of healthcare workers to standard precautions and discussing possible critical issues and preventive strategies. Subsequently, these meetings were scheduled monthly and involved only the NICU staff.
c) Improvement of prevention measures
The following changes in NICU organization and patient management were introduced:
- a new standard protocol for antimicrobial therapy was approved by the Hospital Health Management. The new guidelines standardized the timing for starting and stopping therapy in suspected sepsis, the duration of therapy in confirmed infections and sepsis, and the stopping of therapy after the first negative culture (18);
- hand-washing sensitization posters for caregivers and parents were placed in all rooms;
- contaminated devices were replaced with clean ones after feeding;
- introduction of bundles for common procedures, such as blood-sample collection, diaper change, milk preparation or fridge sanitation.
Evaluation of the Impact of Intervention Measures
The prevalence of MDR-GNB, ESBL-producing GNB, ESBL-KP carriage was assessed and compared in two groups: newborns admitted in the pre-intervention period and newborns admitted in the post-intervention period.
In addition, in order to better evaluate the impact of the intervention measures and identify confounding factors that could impact MDR-GNB carriage, we conducted a quasi-experimental study comparing the clinical features in two subgroups of patients: intervention population and control population.
Inclusion criteria were admission in the NICU and enrollment in the surveillance program for MDR-GNB from November 2017 to March 2018 for the intervention population, and from November 2016 to March 2017 for the control population.
Exclusion criteria were:
- clinical records not available;
- outborn patients with MDR-GNB colonization in the first rectal swab (we do not perform rectal swab on NICU admission so we could not discriminate if MDR-GNB carriage was already present before the admission in the ward);
- patients with a positive rectal swab before the implementation of preventive measures (November 28th 2017).
The outcome variable was detection of MDR-GNB in at least one rectal swab obtained for the surveillance program during the observation period of risk factors exposure.
The observational period was defined in colonized patients as the span of time between NICU admission and the date of the first positive rectal swab, and in non-colonized patients as the span of time between NICU admission and the last rectal swab obtained.
We analysed the following clinical characteristics: type of delivery, sex, gestational age, birth weight, APGAR score at 5’, presence of malformations, feeding (breast milk and/or formulas), use of nasogastric tube, parenteral nutrition, use of invasive devices (central and peripheral venous access), invasive or non-invasive ventilation, surgical treatment, use of antimicrobials and hospital discharge (alive, dead or moved to another hospital).
Collection of Samples and Microbiological Analysis
Collected samples were enriched in liquid cultures (Brain Heart Infusion, Oxoid) for 24 hours at 37°C, then plated in McConkey Agar with three antimicrobial discs (amoxicillin-clavulanate 30μg, meropenem 10μg, ceftazidime 30μg) to detect multi-drug resistant bacteria (13). After overnight incubation, suspected MDR colonies were isolated for identification (standard biochemical methods), susceptibility testing and ESBL detection according with the European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines (19–21).
MDR-GNB were defined as Gram-negative bacteria non-responders in vitro to at least three different classes of antimicrobials under testing (aminopenicillins, third-generation cephalosporins, monobactams, aminoglycosides and carbapenems).
Molecular characterization of MDR-GNB was performed using pulsed-field gel electrophoresis (PFGE) after DNA-cutting with restriction enzymes and electrophoretic profiles were interpreted according to standard procedures (18,22,23).
Statistical Analysis
Prevalence of MDR-GNB carriage, in relation to patient characteristics, procedures and clinical outcomes were compared using the chi-square test or Fisher test for categorical variables and Student’s t-test or the Mann-Whitney U test for continuous variables.
A multivariate analysis by backward stepwise logistic regression was carried out to determine variables significantly associated with MDR-GNB carriage. All variables that did differ between subjects with MDR-GNB carriage (p < 0.10) were initially entered in the model, and the least significant variables were removed one at a time. Goodness of fit of the logistic models was assessed using the Hosmer and Lemeshow test. Several multiple logistic regression models were tested in order to determine the most significant and simplest model with the best available fit for the data.
All significance tests were two-tailed, and p<0.05 was considered significant.
Statistical analysis was carried out by using the R statistical software package (version 3.6.1) and Microsoft Excel 2010.