Five-year evaluation of the antimicrobial susceptibility patterns of bacteria causing bloodstream infections in Iran

Introduction: Bloodstream infections (BSI) are a serious cause of morbidity and mortality worldwide. Emerging antimicrobial drug resistance among bacterial pathogens causing BSI can limit therapeutic options and complicate patient management. Methodology: To encourage the prudent use of appropriate antibiotics in our pediatric population at Children’s Medical Center Hospital, Tehran, Iran, we studied the frequency and antibiogram patterns of blood culture isolates from January 2001 to December 2005. Results: Of 25,223 blood cultures examined, 2,581 (10.23 %) were positive for bacterial growth. The frequency of Gram-positive bacteria isolated was 47.6% (1228 of 2581) and that for Gram-negatives was 52.4% (1353 of 2581). The rates of methicillin (oxacillin) resistance in Staphylococcus aureus and coagulase-negative staphylococci (CoNS) were 79% and 89%, respectively. About 45% of Streptococcus pneumoniae were resistant to trimethoprim-sulfamethoxazole and approximately 66% to penicillin. Among the Gram-negative isolates, Pseudomonas aeruginosa was most frequently isolated, representing 943 (36.7%) over five years. This possibly represents an unrecognized hospital outbreak or contamination of blood culture bottles or other products such as skin disinfectants. Additionally, this pathogen showed extremely high rates of antimicrobial resistance. There were notable differences in frequency of the five most common microorganisms isolated from blood cultures, which can help set priorities for focused infection control efforts. Conclusions: Our findings underscore the need to monitor blood culture isolates and their antimicrobial resistance patterns to observe resistance trends that would influence appropriate empiric treatment and infection control strategies for bacteremic children.


Introduction
Clinically significant bacteremia is a serious consequence of a wide variety of initially localized infections, including those of the urinary tract, respiratory tract, surgical sites, and indwelling devices such as central lines [1].Bloodstream infections (BSI) are an important cause of morbidity and mortality worldwide, affecting over 200,000 individuals annually with a mortality rate of 32% in the United States alone [2][3][4][5].National vital statistics report an increase in age-adjusted death rates due to septicemia from 4.2 per 100,000 in 1980 to 13.2 per 100,000 in 1992, and a number of deaths due to septicemia from 31,224 in 2000 to 33,464 in 2006 in the United States [6].In several European studies, mortality attributed to bacteremias varied from 10% to 50% [7][8][9].
The epidemiology of invasive bloodstream pathogens has changed dramatically over the years [10][11][12].The changes in the incidence and epidemiology of infecting organisms have also brought about an increase in resistance to many antibiotic compounds [10], resulting in a reduction in therapeutic options [11].
Appropriate surveillance by monitoring antimicrobial drug susceptibility trends is a prerequisite to implementing rational measures to tackle the resistance problem [13].In Iran and many other countries, bacterial antimicrobial resistance is a common cause of treatment failure of BSIs.In this report, we describe the frequency, distribution and antimicrobial resistance patterns of blood culture isolates recovered from pediatric patients in Children's Medical Center (CMC) Hospital in Tehran, Iran, from 2001 through 2005.

Study design
This investigation was the second phase of a tenyear comprehensive retrospective study conducted in two five-year phases.The first phase utilized the microbiology laboratory records of 24,600 blood cultures collected from CMC Hospital during 1996-2000 [14].This study represents the second phase of the study and utilized the microbiology laboratory records of 25,223 blood cultures during 2001-2005.In addition to being a referral tertiary care centre, CMC Hospital is one of the educational hospitals of Tehran University of Medical Sciences.This hospital admits patients from all regions of Iran, representing a wide spectrum of socioeconomic levels.

Microbiology data
Microbiology records were reviewed retrospectively.The identification and antimicrobial susceptibility profiles of bacteria isolated from one conventional aerobic blood culture bottle per patient were collected and recorded.The blood culture bottles consisted of trypticase soy or brain-heart infusion broth and were prepared in-house.The bottles were incubated at 35°C for seven days and subcultured to chocolate agar plates within 12 to 24 hours after collection [15,16].Due to annual changes in the hospital formulary, antimicrobial agents utilized and tested varied from year to year.Data regarding the clinical significance of each isolate and whether the BSI was community or hospital-acquired were not available.The antimicrobial agents selected for analysis were those commonly used in the treatment of BSI or septic patients in Iran.Antimicrobial susceptibility results of the isolates were presented as whole numbers that were rounded down at < 0.5 and rounded up at ≥ 0.5.

Organism identification and antimicrobial susceptibility testing
All bacterial isolates were identified in the microbiology laboratory of CMC using standard biochemical identification methods.The biochemical tests included catalase, oxidase, sugar fermentation, bile solubility, coagulase test, motility test, and others routinely used for identification to the genus or species level [15,16].Antimicrobial susceptibility testing was performed by the Kirby-Bauer disk diffusion method, which is the predominant assay utilized in Iran.The guidelines current at the time of the study for interpretation of disk zone diameters as susceptible or resistant followed the National Committee for Clinical Laboratory Standards (NCCLS), which is currently known as the Clinical and Laboratory Standards Institute (CLSI).Due to the high risk of morbidity and mortality among BSI patients, bacteria with intermediate susceptibility results were categorized as resistant to that antimicrobial agent.Quality control bacterial isolates were utilized routinely in the CMC microbiology laboratory to ensure accurate performance of the susceptibility tests [17].

Frequency of occurrence of bloodstream isolates
A total of 25,223 blood culture results from the CMC Hospital in Tehran were analyzed, of which 2,581 (10.23 %) were positive and bacteria were isolated.BSI isolates are displayed in Table 1 by grouping the nine most common bacterial species or groups by frequency of occurrence and year.Gramnegative bacilli were recovered in slightly greater numbers (52.4%; 1,353 of 2,581) than Gram-positive bacteria (47.6%; 1,228 of 2,581).This finding was largely due to a substantial increase in isolates of Pseudomonas aeruginosa during February and March of 2004, resulting in a frequency of 57.9% as opposed to other years which ranged from 13.9 to 34.5% (Table 1).Therefore, the total prevalence of P. aeruginosa was 36.7% of total cases, even surpassing that of coagulase-negative staphylococci (CoNS).This unusual finding may misrepresent the true distribution of pathogen frequency.
The five most frequently isolated sepsis-causing bacteria were as follows: P. aeruginosa, 36.7%;Staphylococcus aureus, 8.4%; Escherichia coli, 5.4% Klebsiella spp.3.5%; and Enterobacter spp.2.6%, collectively accounting for almost 56.6% of the total number of blood culture isolates.The frequency of CoNS and alpha-hemolytic streptococci (except S. pneumoniae) was 31.2% and 3.1%, respectively.These two groups of bacteria probably represent skin contaminants.The frequency of Enterobacter spp.decreased from 6.5% in 2001 to 0.4% in 2005.The frequency of all other pathogens remained essentially stable during the study period (Table 1).

Antimicrobial susceptibility of the most common Gram-positive bacteria
The in vitro susceptibility profiles of the most prevalent Gram-positive bacteria to 10 antimicrobial agents are summarized in Table 2.The rate of methicillin (oxacillin) resistance in CoNS (89%) was higher than that of S. aureus isolates.Only 79% (37 of 178) S. aureus was susceptible to methicillin (oxacillin).Methicillin-resistant S. aureus (MRSA) showed a relatively high rate of resistance to trimethoprim-sulfamethoxazole and gentamicin (≥ 50%) compared to MSSA (Table 2).Only 34% of our S. pneumoniae isolates was susceptible to penicillin.The susceptibility of S. pneumoniae to erythromycin and gentamicin was only 31% and 22%, respectively, whereas susceptibility to clindamycin exceeded 91% (Table 2).Lastly, chloramphenicol showed high activity against the Gram-positive isolates in our study, including staphylococci and S. pneumoniae, with susceptibility rates ranging from 75% to 89% (Table 2).

Antimicrobial susceptibility of Gram-negative bacteria
The antimicrobial susceptibility profiles of the most prevalent Gram-negative bacteria are shown in Table 3. Chloramphenicol was the most active antibiotic against E. coli, K. pneumoniae, and Enterobacter spp., with susceptibility ranging from 53% to 77%.The isolates of P. aeruginosa were extremely resistant to the majority of the antimicrobial agents tested, including ceftazidime (92%).The most active compound against this pathogen was carbenicillin (38% susceptibility).For E. coli, more than 76% of isolates were susceptible to the aminoglycosides (amikacin, gentamicin, tobramycin), cefotaxime, ceftriaxone, and chloramphenicol; however, susceptibility to ampicillin, cephalothin and kanamycin was only 14%, 37% and 45%, respectively.Among the E. coli, ceftazidime resistance, a commonly used phenotypic marker for estimating extended-spectrum betalactamase (ESBL), was 37% (Table 3).
Klebsiella spp. was the third most prevalent Gram-negative isolate.The ESBL rate can be estimated from the low susceptibility to cefotaxime (26%) and ceftazidime (9%).Among the aminoglycosides tested, Klebsiella spp. was most susceptible to amikacin (53%).Chloramphenicol and carbenicillin were the most effective antimicrobial agents against Enterobacter spp., with a susceptibility rate of 57% and 56% respectively, and ampicillin had the lowest susceptibility (9%).

Discussion
In our study, the prevalence of P. aeruginosa was exceptionally high in 2004 when compared to other years studied and possibly represents an Enterococcus spp., with 0.4% frequency, was the 17 th most common BSI pathogen in the current study, in marked contrast to many reports which show enterococci among the top four most frequent BSI pathogens [7,18,19].
Among Gram-positive bacteria recovered from blood cultures, 79% were MRSA, as opposed to much lower rates reported from the USA and Canada [20].In this study, 89% of CoNS were methicillin (oxacillin)-resistant compared to 32% to 37% reported in the SENTRY study, and 68% in Europe [20,21].
However, rates of methicillin-resistant CoNS is similar to those of other reported studies in the United Kingdom (76%) [1], SENTRY 2004 (85%) [22] and the USA (77%) [22,23,24].In contrast to the current data, results from the first five-year phase of the study showed 60% methicillin resistance in S. aureus isolates and 61% among the CoNS isolates [14].Both MRSA and methicillin resistance CoNS isolates were highly susceptible to chloramphenicol, but this antimicrobial is used only rarely in children due to the adverse reaction of aplastic anemia.We suggest that risk-to-benefit studies be undertaken to reconsider the use of chloramphenicol in managing methicillin-resistant staphylococcal pathogens [25].In the current survey, only 14 (34%) of the S. pneumoniae isolates were susceptible to penicillin.In some reports from North America, Latin America, Europe and Egypt, the susceptibility rate to penicillin for S. pneumoniae was 60% to 70% [22,26,27,28] while in the United Kingdom the rate of susceptibility was 100% [1].In a previous report from this center, 44% of S. pneumoniae isolates were susceptible to penicillin.In comparison with our last report and other reports from Iran, we have lower rates of penicillin susceptibility among S. pneumoniae isolates, warranting a possible modification of treatment guidelines for this pathogen in our medical center [14,29].

S. pneumoniae
In the present report, none of the antimicrobial agents tested were sufficiently effective against P. aeruginosa isolates in vitro.Therefore, physicians and hospital formulary groups need to search for other effective antimicrobial agents to treat P. aeruginosa infections [24,30].
E. coli susceptibility against aminoglycosides (77% for gentamicin) is less than the rates cited in many reports from Latin America, the USA, and Canada [23,31,32].The third-generation cephalosporins (ceftazidime and ceftriaxone) showed lower activity than that observed in our previous report [14].Only 9% of Klebsiella spp.and 63% of E. coli isolates were susceptible to ceftazidime which suggests a high prevalence of ESBL-producing strains.This high rate of resistance is rare in other reports (1% in Canada and 2% in the USA) [32].Our isolates were not tested further to confirm ESBL production.
Despite the higher frequency of antimicrobial resistance in Enterobacteriaceae, there are some obvious similarities in patterns: aminoglycosides (amikacin in particular) has the best activity and there are no preferences for third-generation cephalosporins, such as ceftriaxone and ceftazidime.Trimethoprim-sulfamethoxazole, with a 78% susceptibility rate, may be an appropriate treatment for Salmonella infections rather than ceftriaxone, which has a 24% susceptibility rate.In our study, most Gram-negative bloodstream isolates were highly susceptible to chloramphenicol, most probably due to the long time this agent has not been in use.
Our data underscore the need for annual antimicrobial resistance surveillance reports, which can provide valuable insight into resistance trends at a particular medical facility to assist in guidance in the appropriate choice of empiric therapy.

Table 1 .
Frequency of occurrence of bacteria isolated from blood cultures of hospitalized patients 2001-2005 123 unrecognized hospital outbreak or contamination of blood culture bottles or other products such as skin disinfectants.

Table 3 .
In vitro antimicrobial susceptibility of Gram-negative blood culture isolates -Not measured or not reported