Amikacin therapy of patients with multiply antibiotic-resistant Serratia marcescens infections: development of increasing resistance during therapy

Recent independent emergence of multiple multidrug-resistant Serratia marcescens clones within the United Kingdom and Ireland

Serratia marcescens, a member of the Enterobacteriaceae family, is a Gram-negative bacterium responsible for a wide range of nosocomial infections. The emergence of multidrug-resistant strains is an increasing danger to public health. To design effective means to control the dissemination of S. marcescens, an in-depth analysis of the population structure and variation is required. Utilizing whole-genome sequencing, we characterized the population structure and variation, as well as the antimicrobial resistance determinants, of a systematic collection of antimicrobial-resistant S. marcescens associated with bloodstream infections in hospitals across the United Kingdom and Ireland between 2001 and 2011. Our results show that S. marcescens is a diverse species with a high level of genomic variation. However, the collection was largely composed of a limited number of clones that emerged from this diverse background within the past few decades. We identified potential recent transmissions of these clones, within and between hospitals, and showed that they have acquired antimicrobial resistance determinants for different beta-lactams, ciprofloxacin, and tetracyclines on multiple occasions. The expansion of these multidrug-resistant clones suggests that the treatment of S. marcescens infections will become increasingly difficult in the future.

Serratia infections: from military experiments to current practice

Serratia species, in particular Serratia marcescens, are significant human pathogens. S. marcescens has a long and interesting taxonomic, medical experimentation, military experimentation, and human clinical infection history. The organisms in this genus, particularly S. marcescens, were long thought to be nonpathogenic. Because S. marcescens was thought to be a nonpathogen and is usually red pigmented, the U.S. military conducted experiments that attempted to ascertain the spread of this organism released over large areas. In the process, members of both the public and the military were exposed to S. marcescens, and this was uncovered by the press in the 1970s, leading to U.S. congressional hearings. S. marcescens was found to be a certain human pathogen by the mid-1960s. S. marcescens and S. liquefaciens have been isolated as causative agents of numerous outbreaks and opportunistic infections, and the association of these organisms with point sources such as medical devices and various solutions given to hospitalized patients is striking. Serratia species appear to be common environmental organisms, and this helps to explain the large number of nosocomial infections due to these bacteria. Since many nosocomial infections are caused by multiply antibiotic-resistant strains of S. marcescens, this increases the danger to hospitalized patients, and hospital personnel should be vigilant in preventing nosocomial outbreaks due to this organism. S. marcescens, and probably other species in the genus, carries several antibiotic resistance determinants and is also capable of acquiring resistance genes. S. marcescens and S. liquefaciens are usually identified well in the clinical laboratory, but the other species are rare enough that laboratory technologists may not recognize them. 16S rRNA gene sequencing may enable better identification of some of the less common Serratia species.

An outbreak of an unusual strain of Serratia marcescens in two Dublin hospitals

We describe a serious outbreak of infection caused by a strain of Serratia marcescens in two Dublin hospitals which occurred over an 11 week period and affected a total of 15 patients. A contaminated bed-pan macerator in the Intensive Care Unit of one hospital was identified as the possible source of infection and spread of the organism probably occurred via hand transmission by hospital personnel and via patient transfer to a second hospital. All isolates of S. marcescens involved in the outbreak had the same antimicrobial susceptibility pattern, with reduced susceptibility to gentamicin, cefotaxime and ciprofloxacin. Epidemiological typing revealed that the strains of S. marcescens isolated in the outbreak were of an uncommon serotype, O21:K14, and using pulsed-field gel electrophoresis, XbaI DNA macrorestriction profiles clustered at 90% similarity. The DNA patterns of the outbreak strain were also highly similar to S. marcescens isolates of the same serotype recovered from a separate Dublin hospital during the same time period as the outbreak described here. In addition, the isolates clustered at 82% similarity with strains of the same serotype from a retrospective collection of S. marcescens isolates from various hospitals in the Dublin area, indicating that these may be genetic variants of the same strain. Although the outbreak was brought under control following implementation of infection control measures, a significant number of similar O:21 isolates of S. marcescens have since been identified in four Dublin hospitals. These results suggest the unique spread of a single strain of S. marcescens in Dublin hospitals.

Serratia bacteremia

During a 6-yr period, 146 patients at our institution had Serratia bacteremia (3.8% of the total number of episodes of bacteremia), with an incidence of 1.24/1000 admitted patients. We chose a random group of 50 cases for clinical analysis in the present study. The disease was community-acquired in 8% of the cases and nosocomially-acquired in the remaining 92%. The bacteremia was unimicrobial in 84% and part of a polymicrobial bacteremia in 16% of the episodes. The most frequently isolated species of the Serratia genus was S. marcescens. Portals of entry, in decreasing order of frequency, were: urinary, unknown, respiratory, and surgical wound infections. Clinically, the most frequent finding was fever (100%). Shock occurred in 28% of the patients, and none of our cases showed evidence of disseminated intravascular coagulation. We found 62% of Serratia isolates resistant to gentamicin. Overall mortality was 38% and factors associated with a poor prognosis were: severity of the underlying disease, critical clinical situation at onset of bacteremia, presence in the intensive care unit (I.C.U.), occurrence of shock or polymicrobial bacteremia, portal of entry in the respiratory tract, and inadequate treatment.

Development of resistance during antibiotic therapy

The frequency of development of resistance during antibiotic therapy was evaluated by reviewing the literature according to prescribed criteria. Mean resistance rates were calculated to be 9.2% for broad spectrum penicillins, 8.6% for second and third generation cephalosporins, 10.0% for latamoxef, 4.7% for imipenem, 11.8% for ciprofloxacin and 13.4% for aminoglycosides. Emergence of resistance of the infecting organism was associated with therapeutic failure in about half of the cases with the exception of patients treated with aminoglycosides in whom development of resistance resulted in treatment failure in 85% of the cases. The possible benefit of combination therapy in terms of resistance development is discussed.

Review of epidemic aminoglycoside resistance worldwide

Epidemic aminoglycoside resistance may be caused by the spread of a species with distinctive chromosomal genes (e.g., Pseudomonas aeruginosa), or it may be due to the dissemination of plasmids or transposons between genera. Although strains of P. aeruginosa resistant to aminoglycosides because of impermeability may cause nosocomial outbreaks, most of the acute increases in aminoglycoside resistance are due to the spread of inactivating enzymes by plasmids. The index species for intergeneric outbreaks is usually Klebsiella pneumoniae carrying the ANT(2") or AAC(3) gene; however, the distribution of resistance varies greatly by location and species. The AAC(6')-I gene is most common in Serratia marcescens and in East Asian isolates of other species, whereas the AAC(3) gene is common in Chile. In the United States, the ANT(2") and AAC(3) genes are particularly common among Enterobacteriaceae, except for Proteus and Providencia, which often carry the AAC(2') gene. The most common patterns of epidemic resistance lead to the inactivation of gentamicin and, less frequently, tobramycin, but only rarely affect amikacin.

Antibiotic resistance patterns during aminoglycoside restriction

When amikacin first became available its use was restricted to prevent the emergence of resistant strains of gram-negative bacilli to this new agent. Gentamicin was the aminoglycoside most widely used at this time, and the incidence of gentamicin-resistant bacteria was 14%, while only 2.4% were resistant to amikacin. For a period of 15 months gentamicin use was restricted, and amikacin was used almost exclusively. Amikacin use was associated with a fall in the incidence of gentamicin-resistant bacteria to 9.2% (p less than .00005), while amikacin resistance remained unchanged at 2.2% (NS). During a period of 21 months after all aminoglycoside restrictions were lifted, gentamicin use again increased, and was accompanied by a return of gentamicin resistance to the baseline level of 15.3%. During this period, amikacin resistance also increased to 4.0% (p less than .0000001) but was due primarily to an increase in resistant Pseudomonas aeruginosa. Escherichia coli was the most frequently isolated gram-negative bacillus during all three periods, and it remained sensitive to both antibiotics regardless of the drug in use. In contrast, P. aeruginosa showed a high level of resistance to gentamicin, which fell when this antibiotic was restricted, only to return to a high level with reinstitution of gentamicin. While there was also an increase in amikacin resistant strains of P. aeruginosa with unrestricted aminoglycoside use, there was no apparent shift in the pattern of aminoglycoside modifying enzymes among a small random selection of amikacin-resistant bacteria. Impaired uptake of antibiotic was the predominant mechanism responsible for P. aeruginosa resistance among strains that did not produce aminoglycoside acetyltransferase (AAC)(6').

Association of aminoglycoside plasma levels with therapeutic outcome in gram-negative pneumonia

To determine the association of aminoglycoside plasma levels with therapeutic outcome in gram-negative pneumonia, the case reports of 37 patients from four prospective, randomized, controlled trials of gentamicin, tobramycin, and amikacin were analyzed. Twenty (54 percent) of these patients had a favorable outcome. Patients with maximal one-hour postinfusion (peak) levels of 7 micrograms/ml or greater for gentamicin and tobramycin or 28 micrograms/ml or greater for amikacin more often had successful outcomes (14 of 20, 70 percent) than those with levels less than this (six of 19, 32 percent) (p less than 0.006). Patients with overall mean peak levels of 6 micrograms/ml or greater for gentamicin and tobramycin or 24 micrograms/ml or greater for amikacin more often had successful outcomes than those with levels less than this (six of 17, 35 percent) (p less than 0.04). The initial patient temperature, serum urea nitrogen/creatinine ratio, initial polymorphonuclear leukocyte count, and age were also associated with outcome; but by multivariate analysis, achieving an adequate peak concentration was the most important discriminating factor. These results suggest the potential importance of achieving adequate aminoglycoside levels in patients with gram-negative pneumonia.

Infection with netilmicin resistant Serratia marcescens in a special care baby unit

An outbreak of colonisation and infection with a netilmicin resistant strain of Serratia marcescens occurred in a special care baby unit. S marcescens was isolated from a total of 13 babies; significant infection occurred in five, of whom two died. Epidemiological investigation failed to detect a common source but gastrointestinal colonisation of babies formed a prolonged and possibly important reservoir for infection. Containment proved difficult until the unit was closed to new admissions, and even then spread to a temporary unit ensued. O Serotyping and bacteriophage typing disclosed a single epidemic strain. This produced an aminoglycoside acetylating enzyme (AAC(6')) conferring resistance to netilmicin and tobramycin and moderate resistance to amikacin. Use of gentamicin resulted in the isolation of serratia with increased resistance to all aminoglycosides, and, similarly, increased resistance to third generation cephalosporins emerged with their use.

Outbreaks of hospital infection in southwest England caused by gentamicin-resistant Serratia marcescens

Five geographically separate outbreaks of hospital acquired infection caused by gentamicin-resistant strains of Serratia marcescens occurred in the period October 1977 to January 1980 in southwest England. The patients affected were in wards for general or urological surgery, or in neurosurgical, cardiothoracic or general intensive therapy units. Asymptomatic colonization was more common than symptomatic infection, although deaths and serious infections occurred. Control of spread of the bacteria proved difficult. Most strains were resistant to many currently available antibiotics besides gentamicin; only one strain became resistant to amikacin. Representative isolates where characterized by O serotype, bacteriophage type, antibiotic sensitivity pattern, production of beta-lactamases and amino-glycoside-aminocyclitol (ACAG)-modifying enzymes, and plasmid visualization. Plasmid studies provided information that complemented conventional typing methods in determining epidemiological relationships among the outbreaks.

Quantitative antibiogram as a potential tool for epidemiological typing

Forty stock, sink drain and clinical isolates of gram-negative bacilli were tested by broth microdilution against eight aminoglycoside antibiotics. Results show that quantitative antibiograms provide more information as epidemiological tools than qualitative antibiograms.

Clinical, epidemiologic and microbiologic features of a persistent outbreak of amikacin-resistant Serratia marcescens

This article describes a prolonged outbreak (January 1977 to February 1980) of amikacin-resistant Serratia marcescens (ARSM) urinary infections and the methods used for its control. Significant factors predisposing to ARSM urinary tract infection included an extended hospital stay, being in the urology ward, and undergoing urologic surgery. There had been on prior administration of amikacin or of other aminoglycosides in 20 of 27 patients with ARSM urinary tract infections. Chronically infected patients who required multiple hospitalizations represented a major reservoir for the perpetuation of the outbreak, overshadowing the importance of aminoglycoside use. Traditional control measures and even a major change in the inanimate environment were only partially effective in controlling the outbreak, but treatment of bacteriuric patients in the urology unit with "second and third generation" cephalosporins interrupted patient-to-patient transmission. No new cases of ARSM bacteriuria appeared in the urology unit in the ensuing 12 months.

A nosocomial epidemic of antibiotic-resistant Serratia marcescens urinary tract infections

Serratia marcescens is an important pathogen in hospitalized urologic patients. We herein describe an epidemic of 134 urinary tract infections caused by a multipe antibiotic-resistant Serratia marcescens. A common source in the cystoscopy area was responsible for 105 infections Cross-contamination on patient floors amplified the magnitude of the epidemic. There was significant patient morbidity, although no deaths could be attributed directly to the outbreak. Particular attention is directed to patient risk factors and the clinical significance of nosocomial Serratia marcescens infections. The clinical approach to epidemic antibiotic-resistant Serratia urinary tract infection should not rely primarily on antibiotic therapy. Stress is placed on the importance of an interdisciplinary approach to hospital-acquired infections in general and Serratia marcescens urinary tract infections in particular.

Prevalence of gentamicin- and amikacin-resistant bacteria in sink drains

Sink drains from the Veterans Administration Medical Center, University of Oklahoma Health Sciences Center, and the Oklahoma City community were selectively cultured for gentamicin- and amikacin-resistant bacteria. Aminoglycoside-resistant organisms were found in 86% (Veterans Administration Medical Center, 88%; University of Oklahoma Health Sciences Center, 88%; and Oklahoma City community, 77%) of all 233 sink drains sampled. Of 207 sink drains harboring aminoglycoside-resistent organisms, 99% of the organisms were gentamicin resistant and 82% were amikacin resistant. These data suggest that aminoglycoside-resistent organisms are commonly present in the environment.

In vitro susceptibility of gentamicin and/or tobramycin resistant gram-negative bacilli to seven aminoglycosides

The in vitro activity of gentamicin, tobramycin, sisomicin, netilmicin, amikacin, kanamycin and streptomycin was tested simultaneously by the agar dilution method against 584 clinical isolates of gram-negative bacilli that were resistant to gentamicin and/or tobramycin. About half of the gentamicin-resistant Pseudomonas were susceptible to tobramycin but cross-resistance was virtually complete between gentamicin and tobramycin for Enterobacteriaceae. Sisomicin was much more active than gentamicin against Klebsiella, Escherichia and Citrobacter species. Only 18.9%, 27.4% and 27.9% of Klebsiella, Enterobacter and Serratia respectively were resistant to netilmicin. Amikacin was the most effective aminoglycoside with an overall resistance of 15.6%. Kanamycin was effective against 40% of Proteus and Providencia species. Surprisingly, more than half of Klebsiella and Enterobacter species and 85.3% of Serratia species were susceptible to streptomycin.