SURVEILLANCE AND INFECTION CONTROL IN AN INTENSIVE CARE UNIT

Vol. 26 No. 3 INFECTION CONTROL AND HOSPITAL EPIDEMIOLOGY 1 SURVEILLANCE AND INFECTION CONTROL IN AN INTENSIVE CARE UNIT Giovanni Battista Orsi, MD; Massimiliano Raponi, MD; Cristiana Franchi, MD; Monica Rocco, MD; Carlo Mancini, MD; Mario Venditti, MD ABSTRACT OBJECTIVE: To evaluate the effect of an infection control program on the incidence of hospital-acquired infection (HAI) and associated mortality. DESIGN: Prospective study. SETTING: A 2,000-bed, university-affiliated hospital in Italy. PATIENTS: All patients admitted to the general intensive care unit (ICU) for more than 48 hours between January 2000 and December 2001. METHODS: The infection control team (ICT) collected data on the following from all patients: demographics, origin, diagnosis, severity score, underlying diseases, invasive procedures, HAI, isolated microorganisms, and antibiotic susceptibility. INTERVENTIONS: Regular ICT surveillance meetings were held with ICU personnel. Criteria for invasive procedures, particularly central venous catheters (CVCs), were modified. ICU care was restricted to a team of specialist physicians and nurses and ICU antimicrobial therapy policies were modified. RESULTS: Five hundred thirty-seven patients were included in the study (279 during 2000 and 258 in 2001). Between 2000 and 2001, CVC exposure (82.8% vs 71.3%; P <.05) and mechanical ventilation duration (11.2 vs 9.6 days) decreased. The HAI rate decreased from 28.7% in 2000 to 21.3% in 2001 (P <.05). The crude mortality rate decreased from 41.2% in 2000 to 32.9% in 2001 (P <.05). The most commonly isolated microorganisms were nonfermentative gram-negative organisms and staphylococci (particularly MRSA). Mortality was associated with infection (relative risk, 2.11; 95% confidence interval, 1.72 2.59; P <.05). CONCLUSION: Routine surveillance for HAI, coupled with new measures to prevent infections and a revised policy for antimicrobial therapy, was associated with a reduction in ICU HAIs and mortality (Infect Control Hosp Epidemiol 2005;26:000-000). Hospital-acquired infections are a major problem in intensive care units (ICUs). They affect more than 20% of patients in these units, resulting in a mortality rate of greater than 30% 1-8 and additional costs. 9,10 It is widely accepted that surveillance for hospitalacquired infections is needed in ICUs to determine endemic rates, detect outbreaks, and support effective control measures. 1 The scientific basis for ICU surveillance was established by the landmark Study on the Efficacy of Nosocomial Infection Control (SENIC). 11 Moreover, because treatment is usually prescribed before the results of culture and antimicrobial susceptibility testing are available, it is useful for physicians to know the distribution of etiologic agents for particular types of infection and their antimicrobial susceptibilities. An infection control program was created and surveillance was conducted in the ICU of a large teaching hospital in Rome, Italy, to determine the incidence of hospital-acquired infection, related risk factors, microbes causing infection, and antimicrobial susceptibilities. This article reports on the initial effects of this program. METHODS Setting The study was performed in the 13-bed general ICU of the 2,000-bed University Hospital Policlinico Umberto I of Rome, Italy. All patients admitted to the ward for more than 48 hours between January 2000 and December 2001 were prospectively surveyed for hospital-acquired infection. Definitions Centers for Disease Control and Prevention definitions for hospital-acquired infection were used. 12 Only infections occurring more than 48 hours after admission to the ICU were considered ICU acquired. Only the most common ICU-acquired infections were considered: bloodstream infection, pneumonia, urinary tract infection, and surgical-site infection. Drs. Orsi, Raponi, and Mancini are from the Department of Public Health Sciences, University La Sapienza Rome, Rome, Italy. Drs. Franchi and Venditti are from the Division of Infectious Diseases, Internal Medicine Department, Policlinico Umberto I, Rome, Italy. Dr. Rocco is from the Intensive Care Unit, Policlinico Umberto I, Rome, Italy. Address reprint requests to G. B. Orsi, MD, Dipartimento di Sanità Pubblica, Università La Sapienza Roma, P. le Aldo Moro 5, 00185 Roma, Italy. giovanni.orsi@uniroma1.it The authors thank Professor G. M. Fara for his support during the study.

2 INFECTION CONTROL AND HOSPITAL EPIDEMIOLOGY March 2005 TABLE 1 INVASIVE DEVICE MANAGEMENT IN THE INTENSIVE CARE UNIT 2000 2001 P Patients exposed Central venous catheter 82.8% 71.2% <.01 Mechanical ventilation 88.9% 90.5%.49 Urinary catheter 98.9% 98.9%.92 Duration, d* Central venous catheter 11.3 ± 14.5 10.6 ± 10.3.54 Mechanical ventilation 11.2 ± 17.2 9.4 ± 13.5.26 Urinary catheter 12.4 ± 16.0 12.0 ± 14.5.84 *Values are mean ± standard deviation. Data Collection During the 2-year study, an infection control team composed of two physicians specializing in intensive care, two physicians specializing in infectious diseases, and one epidemiologist performed the surveillance. Data were collected prospectively by two specially trained physicians and entered into a database using Epi- Info software (version 2002; Centers for Disease Control and Prevention, Atlanta, GA). 13 The following data were recorded: demographics, dates of admission and discharge, patient origin (ie, emergency department, operating room, ward, or another ICU), admission diagnosis, severity score (SAPS II), underlying diseases (ie, diabetes mellitus, chronic renal failure, cirrhosis, or chronic obstructive pulmonary disease), and final ICU outcome. Initiation and duration were recorded for central venous catheterization, mechanical ventilation, and urinary tract catheterization. 3,8,14-17 All microbiologically or clinically documented infections were recorded, as were all isolated microorganisms and their antibiotic susceptibilities. Samples were taken for culture according to the general principles of specimen collection and transport. 18 Species identification and antimicrobial susceptibility testing were performed on the isolated strains using the VITEK system. Among gram-negative microorganisms, multidrug resistance was defined as resistance to at least five of the following antibiotics: piperacillin, aztreonam, ceftazidime, cefepime, amikacin, gentamicin, ciprofloxacin, imipenem, and meropenem. The incidence of ICU-acquired infections was expressed as the number of infections per 100 patients. In accordance with Centers for Disease Control and Prevention recommendations, we also evaluated the rates per 1,000 patient-days and per 1,000 device-days for device-associated infections. 19 Also, crude (general mortality) and infection-associated (mortality among infected patients) mortality were determined. Prevention Program On the premise that sharing these data might make it possible to influence behavior and reduce the incidence of ICU-acquired infections, 11 there were meetings every 3 months. All epidemiologic data on infection rates were shown to the ICU personnel with emphasis on associated risk factors. The program also included an open discussion on the principal themes emerging from the surveillance and any possible solutions. Based on surveillance results from October 2000 through March 2001, several changes were made to medical and nursing management in the ICU. First, attempts to avoid invasive devices whenever possible (particularly central venous catheters [CVCs]) and to discontinue their use as soon as possible were encouraged to reduce patients exposure to these risk factors. Second, the antimicrobial policy in the ICU was modified to encourage the overall use of beta-lactamase inhibitor combinations (ampicillin sulbactam or amoxicillin clavulanic acid for infections occurring within 48 to 72 hours of ICU admission and piperacillin tazobactam for infections after 72 hours or among patients previously hospitalized in other wards). Use of third-generation cephalosporins was discouraged except for pathogen-targeted therapy. Carbapenems were generally used for lateonset infections (72 to 96 hours after ICU admission) unresponsive to piperacillin tazobactam, and for pathogen-targeted therapy. Finally, to ensure consistent approaches (eg, compliance with hygiene recommendations), only dedicated ICU physicians were allowed to attend patients. Therefore, their total number was reduced from 27 to 19. The study was divided into two periods (2000 and 2001) to evaluate the effect of these changes. Data Analysis Statistical analyses were performed using Epi-Info software (version 2002; Centers for Disease Control and Prevention). Chi-square was used to examine differences between groups. Statistical significance was defined by a P value of less than.05. The univariate relationship between infection and death was tested using relative risk (RR) and its 95% confidence interval (CI 95 ). Multiple logistic regression analysis was used to adjust for potential confounders. RESULTS Sample Characteristics A total of 537 patients were included in the study (279 during 2000 and 258 in 2001). To assess the homogeneity of the two patient clusters, we considered several general characteristics. The two groups (2000 vs 2001) were similar regarding age (54.8 ± 20.3 vs 56.5 ± 19.8 years) and SAPS II score (44.4 ± 17.1 vs 43.6 ± 16.2). Men represented the majority in both years (65.6% vs 63.9%). The primary admission diagnoses were medical (47% vs 48%), surgical (27% vs 33%), and traumatic (18% vs 17%). The most common underlying diseases were heart failure (35% vs 31%) and chronic obstructive pulmonary disease (28% vs 24%).

Vol. 26 No. 3 SURVEILLANCE AND INFECTION CONTROL IN AN ICU 3 TABLE 2 INFECTION RATES IN THE INTENSIVE CARE UNIT DURING THE STUDY PERIOD 2000 2001 2000 vs 2001 No. of Infection No. of Infection Infection Episodes Rate Episodes Rate Reduction P Pneumonia 56 20.1% 44 17.0% -15.4%.37 Bloodstream 51 18.3% 34 13.2% -27.8%.10 Urinary tract 10 3.6% 4 1.5% -58.3%.13 Surgical site 4 1.4% 2 0.8% -42.8%.47 TABLE 3 INFECTIONS PER 1,000 DAYS OF INVASIVE DEVICE USE DURING THE STUDY PERIOD 2000 2001 Standardized Standardized 2000 vs 2001 Infection Rate Rate Reduction P Pneumonia 20.4% 19.3% -5.4%.78 Bloodstream 19.1% 16.6% -13.1%.54 Urinary tract 2.9% 1.3% -58.3%.86 Invasive Procedures Exposure of patients to CVCs declined from 82.8% in 2000 to 71.3% in 2001 (P <.05) and CVC duration was reduced (Table 1). The proportion of patients exposed to mechanical ventilation did not change, but the duration of mechanical ventilation declined from 11.2 days in 2000 to 9.6 days in 2001. Urinary catheterization did not change during the 2 years (Table 1). Infectious Episodes Seventy-one patients (25.4%) were infected before ICU admission in 2000 versus 69 (27.2%) in 2001, with pneumonia accounting for more than 60% of these infections. Overall, 135 (25.1%) of the patients developed 205 ICU-acquired infections, 100 cases of pneumonia (more than 95% occurred in patients receiving mechanical ventilation), 85 bloodstream infections (34.1% were catheter related), 14 urinary tract infections, and 6 surgical-site infections during the study period. Men outnumbered women by 2 to 1 (91 vs 44). The mean age was 54.7 ± 19.9 years, and 42.2% of the patients were older than 65 years. The average SAPS II score was 46.1 ± 14.9. The onset of infection in patients followed ICU admission by a mean of 16.1 ± 15.4 days and a median of 10 days. A total of 80 patients developed 121 ICU-acquired infections in 2000. In 2001, 55 patients developed 84 infections. Therefore, a considerable reduction in the infection rate was observed between 2000 (28.7%) and 2001 (21.3%) (P <.05). The characteristics of ICU-acquired infections and the incidence per 100 patients are provided in Table 2. Rates of ICU-acquired infection per 1,000 days of invasive device use were calculated and showed a smaller decrease between the two periods (Table 3). None of the reductions in individual body-site infections reached statistical significance. In a multiple logistic regression analysis including CVCs, mechanical ventilation, gender, age, and SAPS II score, CVCs (odds ratio [OR], 6.56; 95% confidence interval, 2.74 to 15.70; P <.001) and mechanical ventilation (OR, 5.80; CI 95, 1.34 to 25.18; P =.02) were important predictors of infection. Mortality The results showed a significant reduction in the crude mortality rate between 2000 (41.2%) and 2001 (32.9%) (P <.05). Infection was associated with mortality (RR, 2.11; CI 95, 1.72 to 2.59; P <.05). When the two periods were considered separately, infection remained a significant predictor in both years (2000: RR, 1.98; CI 95, 1.53 to 2.57; P <.05; and 2001: RR, 2.23; CI 95, 1.61 to 3.08; P <.05). Logistic regression analysis showed that CVCs (OR, 3.44; CI 95, 1.65 to 7.17; P <.001), mechanical ventilation (OR, 10.56; CI 95, 1,37 to 81.24; P =.02), and SAPS II score (OR, 3.52; CI 95, 2.17 to 5.71; P <.001) were independent predictors of death. Microorganisms Overall, the most commonly isolated microorganisms during the study period were Pseudomonas aeruginosa and staphylococci (particularly methicillin-resistant Staphylococcus aureus, which accounted for 81% of Staphylococcus aureus infections in 2000 and 86% in 2001). ICU-acquired infection due to gram-negative bacteria was predominant in 2000 (65.9%), particularly nonfermentative species such as P. aeruginosa, Acinetobacter bauman-

4 INFECTION CONTROL AND HOSPITAL EPIDEMIOLOGY March 2005 TABLE 4 DISTRIBUTION OF MICROORGANISMS IN 2000 Bloodstream Urinary Tract Surgical-Site Microorganism Infection Pneumonia Infection Infection Total MRSA 9 (14.1%) 17 (18.1%) - - 26 (14.8%) CNS 14 (21.9%) 2 (2.1%) - - 16 (9.1%) MSSA 3 (4.7%) 3 (3.2%) - - 6 (3.4%) Other gram-positive 4 (6.2%) 6 (6.4%) 1 (9.1%) - 9 (5.1%) Pseudomonas aeruginosa 11 (17.4%) 36 (38.3%) 7 (63.6%) 1 (14.3%) 56 (31.8%) Acinetobacter baumannii 6 (9.4%) 11 (11.7%) - 4 (57.1%) 22 (12.5%) Stenotrophomonas maltophilia 7 (10.9%) 7 (7.4%) - 1 (14.3%) 15 (8.5%) Other gram-negative 7 (10.9%) 12 (12.8%) 3 (27.3%) 1 (14.3%) 23 (13.1%) Candida albicans 3 (4.7%) - - - 3 (1.7%) Total 64 94 11 7 176 MRSA = methicillin-resistant Staphylococcus aureus; CNS = coagulase-negative staphylococci; MSSA = methicillin-susceptible S. aureus. TABLE 5 DISTRIBUTION OF MICROORGANISMS IN 2001 Bloodstream Urinary Tract Surgical-Site Microorganism Infection Pneumonia Infection Infection Total MRSA 9 (20.4%) 15 (21.8%) - - 24 (19.9%) CNS 18 (41.0%) 3 (4.3%) - - 21 (17.3%)* MSSA - 3 (4.3%) - 1 (33.3%) 4 (3.3%) Other gram-positive 3 (6.8%) 5 (7.2%) - - 8 (6.6%) Pseudomonas aeruginosa 5 (11.4%) 26 (37.7%) 1 (20.0%) - 32 (26.5%) Acinetobacter baumannii - 3 (4.3%) - - 3 (2.5%)* Stenotrophomonas maltophilia - 1 (1.5%) - - 1 (0.8%)* Other gram-negative 6 (13.6%) 13 (18.9%) 3 (60.0%) 2 (66.6%) 24 (19.8%) Candida albicans 3 (6.8%) - 1 (20.0%) - 4 (3.3%) Total 44 69 5 3 121 MRSA = methicillin-resistant Staphylococcus aureus; CNS = coagulase-negative staphylococci; MSSA = methicillin-susceptible S. aureus. *P <.05 for comparison with prior period (Table 4). P >.05 for comparison with prior period (Table 4). nii, and Stenotrophomonas maltophilia (Table 4). Perhaps as a result of the changes in hygiene, antibiotic policy, or both, A. baumannii declined from 12.5% to 2.5% and Stenotrophomonas maltophilia from 8.5% to 0.5% (Tables 4 and 5). Multidrug-resistant microorganisms, among all nonfermentative gram-negative bacteria, declined from 29.1% to 8.3%. DISCUSSION SENIC demonstrated that hospital-acquired infection could be reduced by 32% in hospitals where surveillance and prevention programs were implemented. 11 Benefits likely stem from the Hawthorne effect (ie, healthcare workers altering their behavior when watched) and more specific responses to measures addressing particular problems (eg, disinfection of potable water because of a Legionella outbreak). 20-22 Infection control programs need the involvement of all ward personnel to succeed. Success depends on the ability to form a partnership with the ward staff, creating a sense of ownership of the surveillance initiative among them and thus enhancing their cooperation. This was achieved through a partnership with ICU personnel. The study included more than 500 patients. SAPS II scores (> 40) and case mix were typical of a general ICU in a large teaching hospital 23,24 without significant differences between the 2 years, and infection rates before ICU admission did not change. Because preliminary results showed high devicerelated infection rates, reducing exposure to these devices and limiting the duration of their use were emphasized. There was a significant reduction in the infection rate between the two study periods (pneumonia by 15.4%

Vol. 26 No. 3 SURVEILLANCE AND INFECTION CONTROL IN AN ICU 5 and bloodstream infection by 27.8%). Interestingly, infection rates associated with invasive devices confirmed the general reduction, but to a lesser degree. Cases of pneumonia per 1,000 ventilation-days were reduced by 5.4% and bloodstream infections per 1,000 CVC-days diminished by 13.1% (Table 3). The apparent contradictory results may be explained by several factors. First, the reduced number of patients exposed to invasive devices, especially CVCs (P <.01), and the decrease in the duration of exposure may have contributed to the observed reduction. Second, improved personnel education and motivation may have contributed. Finally, it is likely that the Hawthorne effect influenced the behavior of healthcare workers, but it is unlikely that this explained the reduction in 2001 as compared with 2000. The reduction in the crude mortality rate by more than 20% (P <.05) was an important achievement that was likely multifactorial. Preventing infections and making antimicrobial therapy more appropriate to causative pathogens may have contributed to this change. During the first year of study, gram-negative rods caused most infections and A. baumannii was associated with the highest mortality rate (> 70%). The ability of this microorganism to resist many antimicrobial agents makes it especially problematic. 25 Its decrease during the second year of study probably influenced the overall reduction in the crude mortality rate. As expected, P. aeruginosa was the leading gram-negative bacterium and most frequent in pneumonia. During the second year of study, we registered a general decrease in nonfermentative gram-negative bacteria. These bacteria, especially Stenotrophomonas maltophilia, presented alarming levels of multidrug resistance. As observed by others, 26 we suggest that the general change in the microbial flora responsible for infection was possibly affected by the new antimicrobial policy implemented in the ICU. Gram-positive microorganisms caused most bloodstream infections. This did not change during the study period. Most ICU-acquired infections (with both gram-positive and gram-negative bacteria) occurred among patients already undergoing some type of antimicrobial therapy. This suggests the need to reconsider the antibiotic policy of the ICU. The results of this study suggest that surveillance and infection control measures can significantly effect infections and mortality in the ICU. REFERENCES 1. Eggimann P, Pittet D. Infection control in the ICU. Chest 2001;120:2059-2093. 2. Fridkin SK, Welbel SF, Weinstein RA. Magnitude and prevention of nosocomial infections in the intensive care unit. Infect Dis Clin North Am 1997;11:479-496. 3. Jarvis WR, Edwards JR, Culver DH, et al. Nosocomial infection rates in adult and pediatric intensive care units in the United States. Am J Med 1991;91(suppl 3B):185S-191S. 4. Moro ML, Stazi MA, Marasca G, Greco D, Zampieri A. National prevalence survey of hospital-acquired infections in Italy 1983. J Hosp Infect 1986;8:72-85. 5. National Nosocomial Infections Surveillance (NNIS) System. National Nosocomial Infections Surveillance (NNIS) System report: data summary from January 1992-April 2000, issued June 2000. Am J Infect Control 2000;28:429-448. 6. Spencer RC. Epidemiology of infection in ICUs. Intensive Care Med 1994;20:S2-S6. 7. Trilla A. Epidemiology of nosocomial infections in adult intensive care units. Intensive Care Med 1994;20(suppl 3):S1-S4. 8. Vincent JL, Bihari DJ, Suter PM, et al. The prevalence of nosocomial infection in intensive care units in Europe. JAMA 1995;274:639-644. 9. Orsi GB, Di Stefano L, Noah ND. Hospital-acquired, laboratory confirmed bloodstream infection: increased hospital stay and direct costs. Infect Control Hosp Epidemiol 2002;23:190-197. 10. Pittet D, Tarara D, Wenzel RP. Nosocomial bloodstream infection in critically ill patients: excess length of stay, extra costs and attributable mortality. JAMA 1994;271:1598-1601. 11. Haley RW, Culver DH, White JW, et al. The efficacy of infection surveillance and control programs in preventing nosocomial infections in US hospitals. Am J Epidemiol 1985;121:182-205. 12. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC definitions for nosocomial infections. In: Olmsted RN, ed. APIC Infection Control and Applied Epidemiology: Principles and Practice. St. Louis, MO: Mosby; 1996:A1-A20. 13. Franchi C, Venditti M, Pietropaoli P, et al. Hospital infection surveillance by CIN-2000 software in ICU: preliminary results. Presented at the 13th International Conference of the European Society of Intensive Care Medicine; October 1-4, 2000; Rome, Italy. 14. Polderman KH, Girbes AR. Central venous catheter use: Part 2. Infectious complications. Intensive Care Med 2002;28:18-28. 15. Centers for Disease Control and Prevention. Guideline for prevention of nosocomial pneumonia. Respir Care 1994;39:1191-1236. 16. Martin CM, Bookrajian EN. Bacteriuria prevention after indwelling urinary catheterization. Arch Intern Med 1962;110:703-711. 17. Mermel LA. Prevention of intravascular catheter-related infections. Ann Intern Med 2000;132:391-402. 18. Wilson ML. General principles of specimen collection and transport. Clin Infect Dis 1996;22:776-777. 19. Emori TG, Culver DH, Horan TC. National Nosocomial Infections Surveillance System (NNIS): description of surveillance methods. Am J Infect Control 1991;19:19-35. 20. Burke JP. Surveillance, reporting, automation and interventional epidemiology. Infect Control Hosp Epidemiol 2003;24:10-12. 21. Haley RW, Quade D, Freeman HE, et al. Study on the Efficacy of Nosocomial Infection Control (SENIC Project): summary of study design. Am J Epidemiol 1980;111:472-485. 22. Schneeberger PM, Smits MHW, Zick REF, Wille JC. Surveillance as a starting point to reduce surgical-site infection rates in elective orthopaedic surgery. J Hosp Infect 2002;51:179-184. 23. Gattinoni L, Tognoni G, Pesenti A, et al. Effect of prone positioning on the survival of patients with acute respiratory failure. N Engl J Med 2001;345:568-573. 24. Silvestri L, Monti Bragadin C, Milanese M, et al. Are most ICU infections really nosocomial? A prospective observational cohort study in mechanically ventilated patients. J Hosp Infect 1999;42:125-133. 25. Crowe M, Towner KJ, Humphreys H. Clinical and epidemiological features of an outbreak of Acinetobacter infection in an intensive therapy unit. J Med Microbiol 1995;43:55-62. 26. Raymond DP, Pelletier SJ, Crabtree TD, et al. Impact of a rotating empiric antibiotic schedule on infectious mortality in an intensive care unit. Crit Care Med 2001;30:1877-1882.