| Abstract |
Surveillance of hospital-acquired infections (HAIs) in Intensive Care Units is an essential component of infection control programs in hospitals. Critically ill patients tend to have multiple co-morbidities and their treatment involves a variety of medical procedures and invasive techniques, creating potential routes of infection. Over one third of HAIs involve patients in Intensive Care Units, with rates varying from 15% to 40% of admissions, depending on the type of ICU. The commonest types of infection occurring in the ICU setting are bloodstream infections (usually associated with the use of intravascular devices), nosocomial pneumonias (usually ventilator – associated) and urinary tract infections (usually catheter – associated).
The epidemiology of infections in ICUs is best studied with prospective surveillance activities (incidence studies) and the data collected are most appropriately expressed as rates of infection adjusted to the use of invasive devices. In many developed countries the surveillance of HAIs is coordinated at a local or national level with the use of standardized definitions and data collection methodologies, so that intra- or inter-hospital comparisons and benchmarking can be meaningful. For such countries HAIs are becoming an indicator of quality of healthcare services in participating institutions.
In the context of the present study, two prospective, multicenter, incidence surveillance studies of HAIs were conducted in a number of ICUs throughout Greece. The University Hospital of Herakleion was the coordinating hospital for both studies.
16
Study 1. Device-associated nosocomial infection rates in Intensive Care Units in Greece.
Objective: To implement standardized surveillance and estimate site-specific, risk-adjusted incidence rates of intensive care unit – acquired infections.
Design: The study was carried out in polyvalent ICUs in eight hospitals (four of which are university hospitals) in Greece. All the patients who were admitted in the ICUs participated in the study, with prospective surveillance until discharge, using the definitions and protocols of the USA National Nosocomial Infection Surveillance System. For those patients who developed an infection, data on patient characteristics, APACHE II score on admission, date of onset and site of infection and the presence of invasive devices were collected, as well as patient outcome and microorganisms isolated from cultures. The denominator data were collected daily on a different component and included the number of new admissions, total number of patients, number of patients with central venous catheter, urinary catheter, number of patients on mechanical ventilation and patients receiving antibiotics. Surveillance was continued for at least six months in each ICU.
Results: A total of 1,739 patients were admitted to the participating ICUs. Mean length of stay was 10.1 days. Device utilization ratios were 0.95for central venous catheters, 0.81 for mechanical ventilation and 0.98 for urinary catheters. Antibiotics utilization ratio was 0.88. Crude mortality was 21.3%. 501 ICU-acquired infections were detected in 320 (18.4%) patients. The overall infection rate was 28.5 per 1,000 patient-days. Three types of infections accounted for 87.6% of detected infections: Bloodstream infections were the most frequently detected (40.3%), followed by pneumonias (35.3%) and urinary tract infections (12.0%). All infections at these three
17
sites were device-associated. Adjusted to the use of invasive devices, infection rates varied considerably between different ICUs: from 7.5 to 23.1 bloodstream infections per 1,000 CVC-days, from 3.0 to 36.9 pneumonias per 1,000 mechanical ventilation days, and from 0.5 to 7.7 urinary tract infections per 1,000 urinary catheter days. Gram- negative microorganisms were more frequently isolated than gram-positive ones (60.4% and 39.6% respectively), a finding which was constant for all sites of infection. CNS (coagulase negative staphylococci) and Acinetobacter spp were isolated in almost half of primary bloodstream infections. Acinetobacter spp were also the most frequent microorganisms isolated in ventilator-associated pneumonias. Urinary tract infections were most frequently caused by Gram-negative microorganisms such as Pseudomonas spp and Klebsiella spp, while Enterococcus spp and Streptococcus spp were the most frequent gram -positive pathogens isolated.
Conclusions: ICU-acquired incidence infection rates are considerably higher than those published from countries of Western Europe and USA. Device utilization ratios were also high. Implementation of infection control programs in our hospitals including surveillance, benchmarking, comprehensive education and appropriate infection control practices are required in order to reduce infection rates and improve quality of care in Greek ICUs.
Study 2. Surveillance of ventilator-associated pneumonia (VAP) in five Intensive Care Units. Incidence, risk factors, microbiology and patient outcome.
Objective: To design and implement patient-based protocols for targeted surveillance of the incidence of VAP and estimation of risk factors for infection, microbiology and resistance to antibiotics, as well as patient outcome.
18
Design: The study was carried out for a minimum of six months in five intensive care units in Greece. All patients who were admitted during the study period and received mechanical ventilation for more than 48 hours participated in the study with prospective surveillance until discharge (or death). The diagnosis of VAP was made according to the definitions and criteria used by the CDC/NHSN (Centers for Disease Control/ National Healthcare Safety Network). The data collected for every patient enrolled in the study included demographics, the APACHE II score on admission, the duration of mechanical ventilation, the presence of specific pre-determined risk factors for VAP, and patient outcome. In the event of infection, additional data were collected for the isolated pathogens and antibiotic resistance patterns. All data from the participating ICU‟s were entered in a common electronic database for statistical analysis.
Results: Three hundred and eighty-eight patients participated in the study, with 7,154 patient-days in ICU. The mechanical ventilation ratio varied from 0.69 to 0.95 between units. During the study period, 128 (33%) of patients developed VAP. In total, 138 episodes of VAP were reported. The incidence density for all ICU‟s was 17.9 VAP episodes/1,000 patient-days (taking into account the first episode of VAP only) and 27.4 VAP episodes/ 1,000 patient-days at risk. Adjusted to ventilator use, the incidence was 23.6 VAP episodes / 1,000 mechanical ventilation days and 34.9 VAP episodes / 1,000 mechanical ventilation days at risk.
Risk factors that were identified as significant for the development of infection were categorized as patient-related (most significant ones were septic shock and renal failure), intervention-related (of which more significant were thoracic surgery, tracheotomy and re-intubation), and treatment-related (supine position and previous antibiotic use). Gram- negative microorganisms were isolated in 88.2% of infections,
19
most prevalent of which were Acinetobacter spp (39.3%), followed by Pseudomonas spp (24.1%), and Klebsiella spp (11.0%). Of those, 47% of Acinetobacter baumannii isolates were resistant to carbapenems, 31.5% of P. aeruginosa isolates were resistant to piperacillin and 62.5% of Kl. Pneumoniae isolates were resistant to 3rd generation of cephalosporins.
The median length of stay (LOS) for patients with VAP after the occurrence of infection was 12 days, whereas the median total LOS for patients without VAP was 8 days. The mortality between the two groups was not statistically significant.
Conclusions: Within the National Nosocomial Infection Surveillance System of USA (now integrated into the National Healthcare Safety Network), data are collected by using unit- based protocols and are stratified by type of location into 12 ICU specialties. In Greece, were most ICU‟s are polyvalent, infection rates should better be stratified according to patient characteristics, therefore patient-based surveillance protocols should be used, such as those designed for the present study. Surveillance with the use of such protocols is more labor intensive and time consuming (comparing to unit-based protocols) and is better carried out by specially trained, dedicated healthcare professionals.
|
Search
