Ventilator-associated Pneumonia in a Tertiary Care Hospital in India: Role of Multi-drug Resistant Pathogens

Background: Ventilator-Associated Pneumonia (VAP) is the most frequent intensive-care-unit (ICU)-acquired infection. The aetiology of VAP varies with different patient populations and types of ICUs. Methodology: A prospective study was performed over a period of 15 months in a tertiary care hospital to determine the various aetiological agents causing VAP and the prevalence of multidrug resistant (MDR) pathogens. Combination disk method, Modified Hodge test, EDTA disk synergy (EDS) test and AmpC disk test were performed for the detection of extended spectrum beta-lactamases (ESBL), carbapenemases, metallo-beta-lactamases (MBL) and AmpC β-lactamases respectively. Results: Enterobacteriaceae, Haemophilus influenzae, Staphylococcus aureus, Streptococcus pneumoniae, Candida spp. were more common in early-onset VAP, while non-fermenters (Pseudomonas spp. and Acinetobacter spp.) were significantly associated with late-onset VAP (P value 0.0267, Chi-square value 4.91). Thirty-seven (78.7%) of the 47 VAP pathogens were multidrug resistant. ESBL was produced by 50% and 67% of Escherichia coli and Klebsiella pneumoniae respectively. MBL was produced by 20% of P. aeruginosa. AmpC beta-lactamases were produced by 33.3% and 60.7% of the Enterobacteriaceae and non-fermenters respectively. Of the S. aureus isolates, 43% were methicillin resistant. Prior antibiotic therapy and hospitalization of five days or more were independent risk factors for VAP by MDR pathogens. Conclusions: VAP is increasingly associated with MDR pathogens. Production of ESBL, AmpC beta-lactamases and metallo beta-lactamases were responsible for the multi-drug resistance of these pathogens. Increasing prevalence of MDR pathogens in patients with late-onset VAP indicate that appropriate broad-spectrum antibiotics should be used to treat them.


Introduction
Ventilator-associated pneumonia (VAP) is defined as pneumonia occurring more than 48 hours after endotracheal intubation and initiation of mechanical ventilation (MV) including pneumonia developing even after extubation [1].VAP is the most frequent intensive-care-unit (ICU)-acquired infection, occurring in 9 to 24% of patients intubated for longer than 48 hours [2,3].
Early-onset VAP, which occurs during the first four days of MV, usually is less severe, associated with a better prognosis, and is more likely to be caused by antibiotic sensitive bacteria.Late-onset VAP, which develops five or more days after initiation of MV, is caused by multidrug-resistant (MDR) pathogens and is associated with increased morbidity and mortality [4].
A number of studies from India have investigated the causative organisms of VAP.Pseudomonas spp., Acinetobacter spp., Escherichia coli, Klebsiella pneumoniae and Staphylococcus aureus were identified as the common VAP pathogens, with varying prevalence [5][6][7].Up to 40% of these infections can be polymicrobial [3,8].Pseudomonas spp., Acinetobacter spp.and even Enterobacteriaceae are quite often multidrugresistant due to production of extended spectrum beta (β)-lactamases (ESBL), AmpC β-lactamases (AmpC) or metallo-β-lactamases (MBL) [9,10].The aetiological agents of VAP vary with different patient populations and types of ICUs [1,4].Therefore, the local microbial flora causing VAP needs to be studied in each setting to guide more effective and rational utilization of antimicrobial agents.To better understand the aetiology of VAP in India, this study was conducted in two different ICUs in our tertiary care hospital.
The objectives of this study were to determine the prevalence and risk factors of MDR pathogens among our VAP patients and to determine their antibiotic susceptibility pattern as well as detect the presence of ESBL, AmpC β-lactamases, carbapenemases and metallobetalactamases in these VAP pathogens.

Study Design
This prospective observational cohort study was conducted in two intensive care units (ICU) of a tertiary university hospital in India from October 2006 to December 2007.This study was approved by the Research and Ethical committees of Jawaharlal Institute of Postgraduate Medical Education and Research (JIPMER) and informed consent was obtained from each patient's next of kin.

Setting
The study was conducted in the Medicine Intensive Care Unit (MICU) and Critical Care Unit (CCU) of Jawaharlal Institute of Post-Graduate Medical Education and Research (JIPMER).This is an 860-bed tertiary care hospital and Institution of National Importance in India.The departments of Microbiology, Medicine and Anesthesiology and Critical Care were involved in this study.Each ICU is comprised of 8 well-spaced beds and patients were either admitted directly to the ICU or transferred from other wards, namely medicine, surgery, obstetrics, neurology and cardiology wards.Postoperative patients requiring ventilation were admitted in the CCU, while the patients with medical conditions necessitating ventilation were admitted in the MICU.Three nurses are posted in an ICU with a nurse patient ratio of 1: 2.7.

Subject and sample size
During the 15-month study period, a total of 538 patients who were intubated and on mechanical ventilation in the CCU and MICU were prospectively reviewed.Among them only 206 patients who were ventilated for more than 48 hours were eligible for inclusion in the study.Six of these 206 patients were assessed to have developed pneumonia within 48 hours of mechanical ventilation and were excluded.
The remaining 200 patients were included in the study.

Procedure for data collection
All patients included in the study were monitored at frequent intervals (every three days) for the development of VAP using clinical and microbiological criteria until either discharge or death.The clinical parameters were recorded from their medical records and bedside charts.Details of antibiotic therapy, surgery, use of steroids, duration of hospitalization, presence of neurological disorders, and impairment of consciousness were also noted.

Criteria for diagnosis of VAP
The diagnosis of VAP was based on clinical and microbiological criteria [3].A clinical suspicion of VAP was made in patients with a Modified Clinical Pulmonary Infection Score (CPIS) > 6 [11]; the diagnosis was confirmed by performing a quantitative culture of the endotracheal aspirate and observing ≥ 10 5 cfu/ ml [12][13][14].Based on these criteria, 36 of 200 enrolled patients were diagnosed with VAP.

Microbiological techniques
The organisms isolated by quantitative culture of the endotracheal aspirate (EA) from VAP patients were identified based on standard microbiological techniques [15].The susceptibility of the clinical isolates to some routinely used antibiotics was determined by the Kirby-Bauer disk diffusion method [16].
Combination disk method using both cefotaxime and ceftazidime, alone and in combination with clavulanic acid, was performed for detection of extended spectrum -lactamase (ESBL) among the members of Enterobacteriaceae [18].Five mm or more increase in zone of inhibition for either cefotaxime-clavulanic acid or ceftazidime-clavulanic acid disk compared to the cefotaxime or ceftazidime disk respectively was taken as confirmatory evidence of ESBL production.Amp C disk test was performed for detection of AmpC -lactamase [19].A flattening or indentation of the cefoxitin inhibition zone in the vicinity of the disk with test strain was interpreted as positive for the production of AmpC β-lactamase.An undistorted zone was considered as negative.Modified Hodge test was carried out for detection of carbapenemase as described previously [20].The presence of a cloverleaf-shaped zone of inhibition due to carbapenemase production by the test strain was considered positive.EDTA disk synergy test (EDS) was done using both meropenem and ceftazidime for detection of metallo-β-lactamases (MBL) [20].The presence of an expanded growth inhibition zone between meropenem or ceftazidime and EDTA was interpreted as positive for MBL production.VAP pathogens, such as Pseudomonas spp., Acinetobacter spp., and enteric Gram-negative bacilli expressing ESBL, AmpC β-lactamases or MBL, MRSA and multidrug-resistant S. pneumoniae (resistant to penicillin and at least two other antibiotic classes) were defined as "multi-drug resistant" (MDR) pathogens [4,21].

Statistical analysis
Data entry and analysis were done using SPSS for Windows Version SPSS 16.0 (SPSS Inc, Chicago, Illinois).Means and standard deviations (SD) were calculated as required for numerical variables.
The chi-square test or Fisher's exact test was used to compare two groups.Univariate analysis was used to compare the variables for the outcome groups of interest.We confirmed the results of these tests with logistic regression analysis.This was necessary to avoid producing spuriously significant results with multiple comparisons.Results of the logistic regression analyses are reported as adjusted odd ratios with their 95% confidence intervals.All P values < 0.05 were considered statistically significant.

Results
The demographic data of the 36 patients diagnosed with VAP have been described in detail in our previously published article [3].

Early-onset VAP pathogens versus late-onset VAP pathogens
Enterobacteriaceae, H. influenzae, S. aureus, S. pneumoniae, Candida spp.were more common in early-onset VAP, while non-fermenters were significantly associated with late-onset VAP (P value 0.0267, Chi-square value 4.91) (Tables 1 and 2).The antibiotic resistance pattern of the various etiological agents of early-onset VAP and late-onset VAP are summarized in Tables 1 and 2. None of the Acinetobacter spp.causing early-onset VAP were colistin resistant, while 20% resistance to colistin was observed among Acinetobacter spp.associated with late-onset VAP.

Comparison of bacterial patterns of VAP in MICU and CCU
Non-fermenters (77.8%) were the most predominant pathogens causing VAP in the CCU, while in the MICU along with non-fermenters (48.3%), members of Enterobacteriaceae (24.1%) and Gram-positive bacteria (24.1%) commonly caused VAP.VAP episodes due to Gram-positive bacteria (5.6%) were relatively less common in the CCU (Figure 1).

Detection of ESBL, AmpC β-lactamase and Metallobetalactamase
ESBL was produced by 50% and 67% of E. coli and K. pneumoniae respectively.Only two of the ten P. aeruginosa isolates were tested for metallobetalactamases and both were positive, while AmpC β-lactamases were produced by 33.3% and 60.7% of the members of Enterobacteriaceae and non-fermenters respectively (Table 3).

MDR pathogens
Thirty-seven (78.7%) of the 47 VAP pathogens in our study were multi-drug resistant (MDR).These MDR pathogens included Gram-negative bacteria (non-fermenters and members of Enterobacteriaceae) producing ESBL, AmpC βlactamases or MBL, MRSA and S. pneumoniae showing resistance to oxacillin, tetracycline and erythromycin.

Risk Factors
Administration of prior antibiotic therapy was a significant risk factor for VAP caused by MDR pathogens (RR, 2.18; 95% CI, 0.74 to 6.42; P = 0.0404).Table 4 shows that current hospitalization of five days or more was also a significant risk factor for VAP caused by MDR pathogens (RR, 1.79; 95% CI, 0.88 to 3.61; P = 0.0301).Multivariate logistic regression analysis confirmed that prior antibiotic therapy and current hospitalization of five days or more were independent predictors of VAP caused by MDR pathogens (Table 5).

Discussion
VAP is an important nosocomial infection among ICU patients receiving MV.Multidrug resistant pathogens such as P. aeruginosa, A. baumannii and S. aureus (42.9% of them being MRSA) were the common organisms causing VAP.This highlights the need for treatment of the VAP cases with second-line antibiotics effective against these MDR pathogens.This finding also emphasises the need for stringent preventive measures for VAP, as the treatment of an established VAP becomes very expensive [22].Nonfermenters such as Pseudomonas spp.and Acinetobacter spp.were significantly associated with late-onset VAP as it was observed by other workers [23,24].But in our study even in patients with earlyonset VAP, Acinetobacter spp. was the most common pathogen because most of them had risk factors for MDR pathogens.
Late-onset VAP was associated with higher rates of infection with MRSA and colistin resistant MDR Acinetobacter spp., but the resistance of the nonfermenters to the other antibiotics was almost the same in both early-and late-onset VAP.Many of the early-onset VAP cases had the risk factors such as prior antibiotic therapy and current hospitalization for five days or more for infection with MDR pathogens.That could be the reason for the almost similar susceptibility pattern of the isolates from late-onset and early-onset VAP.Even the American Thoracic Society guidelines supports the same reasoning by suggesting that patients with early-onset VAP who have received prior antibiotics or who have had prior VAP -Ventilator-associated pneumonia AMP -Ampicillin, TIC -Ticarcillin, PTZ -Piperacillin-tazobactam, GEN -Gentamicin, AMK -Amikacin, CIP -Ciprofloxacin, GAT -Gatifloxacin, CTR -Ceftriaxone, CAZ -Ceftazidime, MEM -Meropenem, CL -Colistin, PEN -Penicillin, OXA -Oxacillin, CFL -Cephalexin, TET -Tetracycline, ERY -Erythromycin, VAN -Vancomycin MSSA -Methicillin sensitive Staphylococcus aureus MRSA -Methicillin resistant Staphylococcus aureus hospitalization within the past 90 days are at greater risk for colonization and infection with MDR pathogens and should be treated similarly to patients with late-onset VAP [4].
We also observed that non-fermenters (77.8%) were the most predominant pathogens causing VAP in the CCU, while in the MICU along with nonfermenters (48.3%), members of Enterobacteriaceae (24.1%) and Gram-positive bacteria (24.1%) were also commonly causing VAP.VAP episodes due to Gram-positive bacteria (5.6%) were relatively less common in the CCU.The knowledge of this difference in pathogens causing VAP in different ICU settings will guide the administration of appropriate empirical antibiotics for treatment of the infection.
We observed that colistin is highly active against Acinetobacter spp., while piperacillin-tazobactam has good activity against Pseudomonas spp.But as we have studied only a small number of isolates, these findings need to be further confirmed by larger clinical trials, as they may have a major impact on the treatment of these VAP pathogens.AmpC β-    [25,26].Similarly ESBL and AmpC β-lactamases were produced by a large proportion of the Enterobacteriaceae.
In the present study, we found that prior antibiotic therapy and current hospitalization of five days or more were independent predictors of VAP caused by MDR pathogens by multivariate logistic regression.This emphasizes the need for judicious selection of patients for antibiotic therapy.The prophylactic use of antibiotics is therefore not recommended, and exposure to antibiotics is a significant risk factor for colonization and infection with nosocomial multidrug-resistant pathogens as observed by other authors [1,27,28].The rational use of appropriate antibiotics may reduce patient colonization and subsequent VAP with MDR pathogens.Similarly, unnecessary prolonged hospitalization of the patients should be avoided as far as possible.But it may not be feasible in most situations as the patients' condition may demand prolonged hospital stay.However, the knowledge of this risk factor should suggest the possibility of infection due to MDR pathogens in patients developing VAP after hospitalization for five days or more.
As the study was conducted in a resource-limited setting, only small number of patients with VAP in a single center were studied, which could be considered a limitation of our study.In addition, we recognize that the findings of this study may not necessarily reflect the situations in other similar centers in India.Hence, we suggest further multicentered studies with larger patient numbers to confirm our findings, in particular the high incidence of MDR pathogens.
To conclude, VAP is increasingly associated with MDR pathogens.Production of ESBL, AmpC βlactamases and metallo β-lactamases were responsible for the multi-drug resistance of these pathogens.Knowledge of the susceptibility pattern of the local pathogens should guide the choice of antibiotics, in addition to the likelihood of organisms (early-or late-onset VAP).As there was an increasing prevalence of MDR pathogens in lateonset VAP, appropriate antibiotics should be used to treat them.Patients with early-onset VAP who have received prior antibiotics or who were hospitalized earlier should also be treated similarly to those with late-onset VAP, as they are at higher risk for infection with MDR pathogens.Colistin and piperacillin-tazobactam may be used for successful treatment of multi-drug resistant Acinetobacter spp.and Pseudomonas spp.respectively as they showed good in vitro activity against these MDR pathogens.

Table 4 .
Univariate analysis of the risk factors for VAP by MDR pathogens.

Table 5 .
Logistic regression analysis of the risk factors for VAP by MDR pathogens.P