Nosocomial infections in a neonatal intensive care unit during 16 years: 1997-2012

INTRODUCTION

Newborns admitted to intensive care units (ICUs) are at high risk for developing nosocomial infections (NIs) because of the severity of their illness and exposure to invasive medical devices such as mechanical ventilators and central venous catheters (CVCs) and resistant microorganisms^1,2.

Surveillance of NIs is an essential part of quality patient care; however, there are few reports of National Healthcare Safety Network (NHSN) surveillance in neonatal intensive care units (NICUs) and none in developing countries³. The objective of this study was to report the incidence of NIs, causative organisms, and antimicrobial susceptibility patterns in a large cohort of neonates admitted to the NICU during a 16-year period.

METHODS

The Medical Hospital of the Federal University of Uberlândia is a 533-bed public teaching general hospital and a tertiary care center. The NICU consists of 2 rooms with a capacity of 15 neonates and also serves as a referral center for several hospitals in the vicinity.

From January 1997 to December 2012, all neonates admitted for more than 24h were included. Participation in the study was voluntary and confidential. The patients included were followed 5 times per week from birth to discharge or death. Epidemiological surveillance was conducted according to the National Nosocomial Infections Surveillance System⁴ and National Healthcare Safety Network (NHSN)⁵. In addition to standard hospital epidemiological measures, we also analyzed data adjusted for the duration of exposure to the hospital environment or to invasive devices (reported as patient-days and device-days, respectively). The neonates were stratified into the following birth weight categories: ≤ 750; 751-1,000; 1,001-1,500; 1,501-2,500 and > 2,500g.

An nosocomial infection was defined as an infection not present or incubating at the time of NICU admission, with onset after 48h of stay. Mechanical ventilator-associated pneumonia (VAP) was defined as the development of pneumonia in a neonate 48h or more after initiation of mechanical ventilation. Primary bloodstream infection (pBSI) was defined as laboratory-confirmed bloodstream infection (BSI) or clinical sepsis not related to infection at another site, except for an intravascular device. A central venous catheter-related pBSI was defined as the presence of all of the following criteria: I) bacteremia or fungemia in a neonate with a CVC with at least 1 positive blood culture obtained from a peripheral vein; II) clinical manifestations of infection (i.e., fever [>38°C], hypothermia [<37°C], apnea, or bradycardia) and no apparent source for the BSI except the CVC; and III) a positive semiquantitative (>15 colony-forming units [CFU]/catheter segment) or quantitative (>10³ CFU/catheter segment) culture from which the same organism (species and antibiogram) was isolated from the catheter segment and peripheral blood. Umbilical catheters were considered central lines⁶.

In accordance with institutional protocols, the same initial diagnostic work-up was applied to all neonates upon suspicion of pBSI or VAP, irrespective of the clinical condition. This included a plan of clinical assessment and care concerning thermoregulation, ventilation, circulation, metabolism, and nutrition. All infants underwent a complete physical examination, continuous electrocardiography, pulse oximetry, and central or peripheral venous catheterization as needed. A chest radiograph was performed, and blood samples were drawn for arterial gasometry; a complete hemogram; and sodium, potassium, chlorine, glucose, and C-reactive protein measurement. Two blood cultures were drawn on separate occasions before initiation of antimicrobial treatment in all neonates who had suspected pBSI or VAP, except for extremely low birth weight neonates or neonates who had fulminant sepsis, in whom antimicrobial treatment was initiated immediately after a single blood culture was drawn. Cultures for anaerobic bacteria were not routinely obtained.

Blood cultures that were positive for organisms that generally are considered to be contaminants (e.g., Corynebacterium sp., Bacillus sp., micrococci, and diphtheroids) were excluded from analysis. The coagulase-negative staphylococci (CoNS) were considered true pathogens when 2 blood cultures drawn on separate occasions were positive for the same species and antibiogram and the organism was not related to infection at another site or when 1 blood culture was positive in a patient with an intravascular access device and the physician instituted appropriate antimicrobial treatment. For all other pathogens, a single positive blood culture was considered significant.

Specimens for culture were collected based on clinical criteria established by the medical staff. Microbiological data were obtained from the Uberlândia University Hospital Microbiology Laboratory.

RESULTS

From January 1997 to December 2012, 4,615 neonates were admitted to the NICU, and 192 months of data representing 62,412 patient-days were available for analysis.

The average weight of the neonates was 1,830.6g, with the highest proportion (34%) of neonates in the weight category >2,500g. The proportion of neonates with a gestational age >31 weeks was 55.3%. The average length of stay in the unit was 14.9 days. The invasive device use rate [number of device-days/number of patient-days (times) 100] was 0.8 for CVC and 0.4 for mechanical ventilation. The device-associated infection rates were 17.3 pBSIs per 1,000 central line-days and 3.2 pneumonia infections per 1,000 ventilator-days. The mortality rate in neonates with NI was 10.4%, and overall mortality was 9.1% (Table 1).

In total, 902 (19%) infants developed an NI, and 194 (21.5%) of these had more than 1 infection for a total of 1,096 NIs, principally BSI (752; 68.6%) and conjunctivitis (208; 19%) (Table 2).

A total of 1,182 microorganisms were isolated from sterile body site cultures in 902 neonates. Gram-positive bacteria and gram-negative bacteria accounted for 636 (53.8%) and 460 (38.9%) isolates, respectively. Fungi accounted for the remaining 86 (7.3%) isolates. The distribution of microorganisms isolated according to year is shown in Figure 1.

FIGURE 1 –  Distribution of microorganisms in 2-year intervals. GPB: gram-positive bacteria; GNB: gram-negative bacteria. 

In a comparative analysis of infection episodes occurring during the 16-year study period, we observed a reduction in the years 2004 and 2005 owing to a retirement unit and stricter hand hygiene and infection control measures. The distribution of episodes according to year is shown in Figure 2.

FIGURE 2 –  Distribution of episodes per 1,000 patient-days during 1997-2012.  

CoNS (34.3%) and Staphylococcus aureus (15.6%) were the most common etiologic agents isolated from cultures. Selected antimicrobial susceptibility patterns are presented in Table 3, showing that 86.4% of CoNS and 28.3% of Staphylococcus aureus were oxacillin-resistant. Of special interest is the resistance of Serratia marcescens, Klebsiella pneumoniae, and Pseudomonas aeruginosa to third-generation cephalosporin, the first 2 in BSI cases and the last in conjunctivitis cases.

DISCUSSION

Analysis of our data suggests that participation in surveillance systems, like the NHSN, is important for comparing individual NICU data with other data from the literature. However, comparing surveillance data from different countries is not straightforward. For example, device use rates vary widely between different countries, suggesting differences in either illness severity or clinical practice.

There is little information in the literature regarding the usefulness of assessing infection rates according to patient-days or device-days, that is, standard hospital epidemiological measures adjusted for the duration of exposure to the hospital environment or to invasive devices, such as intravascular catheters or mechanical ventilation, in large cohorts of neonates admitted to NICUs in developing countries⁷. To the best of our knowledge, this study represents the largest cohort to date that evaluates this issue in the developing world. As in most previous reports, BSI was the main cause of NI. In this study, BSI occurred at a rate of 12.0 per 1,000 patient-days and accounted for 16.3% of NIs. A higher proportion (17.3 per 1,000 CVC-days) of all BSIs could be linked etiologically to a CVC. Other authors have reported a BSI incidence between 15.6% and 32.6%⁸. In our study, the rate of ventilator-associated pneumonia (VAP) was 3.2 per 1,000 ventilator-days. According to the NHSN, median VAP rates are 2.0 per 1,000 ventilator-days⁹. Because many NHSN hospitals use only microbiological reports to identify patients with VAP, NHSN rates may underestimate the true incidence of VAP. Therefore, the current VAP rate is probably more representative, with all diagnoses of VAP confirmed based on the review of patient medical records and microorganisms recovered from tracheal aspirates on the day of diagnosis.

Gram-negative rods are major pathogens of neonatal sepsis in developing countries¹⁰. In NICUs, following the adoption of sophisticated tertiary neonatal care with a high rate of invasive device use, CoNS stand out as the main agents of neonatal nosocomial sepsis, with Staphylococcus epidermidis the most frequently isolated¹¹. Neonatal infection by CoNS is less severe but causes significant morbidity especially among infants of very low birth weight. In this study, CoNS was the most common microorganism causing NIs with microbiological diagnoses. These results are in strong agreement with other studies, similar to previous studies conducted in the unit^12,13.

During this study, we observed a gradual decrease in gram-negative and an increase in gram-positive bacterial infection rates. In particular, in 2004, there was a reduction in infection rates demonstrated by a study conducted in the NICU to evaluate the impact of the NICU building environment on the risk of NI by surveillance during a 4-year period during which the NICU was moved from an old building to a temporary unit and finally to a new and better-designed unit, all in the absence of a change in staffing¹⁴.

Most CoNS isolates recovered from infants in this study were resistant to oxacillin. Nevertheless, similar to findings reported by Celebi et al.¹, clinical improvement was documented in more than half of the patients. The CoNS isolates represented true bacteremia, and all study patients met the clinical criteria for sepsis. One possible explanation is that CoNS is a low-virulence pathogen. According to Ben Jaballah et al.¹⁵, high resistance to third-generation cephalosporin by gram-negative bacilli has increased, includingAcinetobacter sp, Enterobacter sp, Serratia marcescens, Klebsiella pneumoniae, and Pseudomonas aeruginosa. This was shown in some of our studies including detection of outbreaks of these pathogens^16–18.

The most important NI remains BSI with staphylococci as the predominant pathogens, which showed much higher rates than those reported in the literature¹⁹. Multiresistant microorganisms, especially oxacillin-resistant staphylococci and gram-negative bacilli resistant to cephalosporin were frequently found. The process of evaluating the causative organisms was valuable in itself by helping to draw the attention of health care professionals to this important cause of morbidity and by promoting the use of strict hygiene measures and meticulous care of the infected infants.