Journal
of Hospital
Infection
(1991)
Epidemiological marcescens
study of an outbreak in a haemodialysis
P. U. Krishnan, Departments
18, 57-61
B. Pereira*
of Serratia unit
and R. Macaden
of Microbiology and *i?ephrology, St. John’s and Hospital, Bangalore 560034, India Accepted for publication
14 January
Medical
College
1991
Summary:
Twenty-three isolates of Serratia mamescens were isolated over a lo-month period from the blood and arteriovenous shunt sites of patients undergoing haemodialysis in an artificial kidney unit. Surveillance measures performed-on the equipment, sterile materials! environment and personnel of this unit vielded Serratia from the air condmoner and one of the dialvsis units. The isolates from the patients and dialyser unit were pigmented &d had an identical biochemical profile, antibiograms, phage typing pattern and 0 serotype. The isolate from the air conditioner, though of the same biotype, had a different phage and serotype. It was concluded that the dialyser was the ‘common source’ and that the organism was persisting in the machine in spite of recommended sterilization procedures being implemented. Keywords:
Sevratiu
marcescens; haemodialysis
unit;
epidemiological
survey.
Introduction Hospital-acquired infections by Serratia marccscens have been well documented in the immunocompromised host.‘,’ Over the past two decades, outbreaks of septicaemia, meningitis and wound infections have been increasingly reported, especially in patients subjected to invasive diagnostic and therapeutic procedures. 3*4The morbidity and mortality rates in these infections are generally high since most patients have serious underlying medical problems and the infecting strains are usually multiply drug resistant.4 Epidemiological marker studies have been done in an attempt to trace the source of infection in some of these outbreaks.4,5 Various types of outbreaks occur in the Haemodialysis Unit and bacterial infections-especially with hospital-acquired pathogens-are recognized to be the greatest hazard.6 The following report details the clinical and bacteriological features of an outbreak of serratia septicaemia in a Haemodialysis Unit over a lo-month period.
Correspondence 019s4701/91/050057+05
to: Dr P. U. Krishnan to3.00/0
0 1991 The Hospital
57
Infection
Society
58
P. U. Krishnan Materials
et al.
and methods
St. John’s Medical College Hospital is a 750-bed teaching hospital in South India. It has an artificial kidney unit which carries out about 150 haemodialysis procedures per month. Routine bacteriological survey of the dialysis fluid, re-use tubes and water treatment system is carried out once a month according to standard recommendations.7 The dialyser is re-used two or three times, after being disinfected by recommended techniques.6 Three cases of S. marcescens septicaemia were observed over a period of 2 weeks and these were brought to the notice of the Infection Control Committee. Surveillance measures were undertaken and samples were taken for culture from the water treatment system, haemodialysis fluid concentrate, re-use tubes, all equipment and furniture in the dialysis room and the air conditioner vents. The unit was temporarily closed for thorough disinfection and fumigation. Subsequently, on re-opening of the unit, three further cases of serratia infection either of blood and/or the arteriovenous shunt site occurred in the same unit over a period of 3 months, after which there were no further cases for 2 months. After this period and over the following 4 months, 15 sporadic cases of serratia septicaemia occurred in the dialysis unit. A second surveillance similar to the first was done. In addition, to test the bactericidal effect of 2% formalin (used for dialyser disinfection), a broth culture of the serratia isolate was kept in contact with this solution for 48 h at 37°C and then subcultured. Nail clippings and finger-web swabs were taken from Dialysis Unit personnel to detect carriers. The case records of all these patients were reviewed and analysed. All culture techniques, bacterial identification and antibiotic susceptibility tests were done according to recommended procedures.‘,” All the serratia isolates were sent to the Public Health Laboratory Service, London with a request to Dr T. L. Pitt for epidemiological marker studies. Results
In all, 26 isolates of S. maxescens were obtained, 21 from blood, two from vascular access sites, one from the environment (air conditioner vent), one from the dialysing machine No.1 and one from the contents of the dialyser kept ready for re-use. None of the hospital personnel were found to be carriers. The 2% formalin was found to be bactericidal for these serratia isolates. All the strains were pigmented, producing the typical cherry-red pigment prodigiosin. The biochemical profile of all strains, including the environmental one and those from the equipment, was identical. All isolates were susceptible to streptomycin (10 pg), choramphenicol (30 pg), kanamycin (30 pg), gentamicin (10 pg), cotrimoxazole (2.5 pg) and resistant to tetracycline (3Opg) and ampicillin (1Opg).
Serratia
marcescens
59
outbreak
All except the isolate from the air conditioner belonged to the 06 serological group (Table I) and had the same phage typing pattern isolate belonged to the 08 group and was (1/3/4/7/9/l 1). Th e air conditioner of a different phage type (l/3/5/7) All the patients had end stage renal failure and were undergoing dialysis in this Unit. In two of them Serratia was also isolated from the vascular access sites. Blood cultures were done in the first two cases when these patients presented with low-grade fever without localizing signs following haemodialysis. As the outbreak became established, blood cultures were done as a screening measure. S. marcescens was isolated from the blood cultures of 21 patients. In addition, swabs from the arteriovenous shunts of two of these patients yielded Serratia. Except for one patient who died of a massive cerebral infarction while on treatment, and another who developed bacterial endocarditis with the same organism, all the others had an uneventful recovery either on oral or parenteral antibiotic therapy. Discussion The incidence of S. marcescens infection in our hospital has never exceeded five cases per year over the last five years, and even among these, episodes of significant bacteraemia have been rare. This sudden and clustered increase in the isolation of Serratia from blood was a cause of alarm and various measures were taken to curb this outbreak. Fortunately, unlike previously described outbreaks, i,ll this strain was susceptible to commonly used antibiotics. Farmer et aZ.‘* have repor t ed that the initial resistance may usually be R-factor mediated and often includes resistance to the more useful, less toxic drugs such as tetracyclines, sulphonamides and ampicillin. Table
1. Analysis
Source
Blood
of Serratia
No. of isolates
culture
marcescens
isolates from patients
Antibiogram*
and environment
Serotype
Susceptibility pattern
Resistance pattern
Phage type
21
S,C,K,G,Cot
T,A
06
l/3/4/7/9/1
1
Vascular access site cultures
2
S,C,K,G,Cot
T,A
06
l/3/4/7/9/1
1
Air conditioner vent
1
S,C,K,G,Cot
T,A
08
l/3/5/7
Dialysing machine
1
S,C,K,G,Cot
T,A
06
l/3/4/7/9/1
1
1
S,C,K,G,Cot
T,A
06
l/3/4/7/9/1
1
Contents of dialyser
No. 1
*S, Streptomycin; C, chloramphenicol; tetracycline; A, ampicillin.
K, kanamycin;
G, gentamicin;
Cot,
cotrimoxazole;
T,
60
P. U. Krishnan
et al.
If these organisms are not eliminated, a vicious circle of increasing drug resistance will continue until a sub-strain resistant to all antibiotics emerges. In our hospital, however, even though the outbreak has continued for 10 months, multiply drug resistant strains have not emerged as yet. The source of infection has in previous outbreaks been traced to gastrointestinal carriage, especially in infants,” contaminated handwashing brushes,’ shaving brushes, l3 barber’s hands4 and inhalation therapy medications.14 However, it has been suggested that environmental reservoirs are insignificant as a source. l&l6 Initially, the air-conditioner was implicated as a common source because the isolate from that site had the same biochemical properties and antibiotic sensitivity pattern as the clinical isolates. Also, with the advent of the monsoon and the decreased use of the air-conditioner, the outbreak subsided. However, after a lapse of 2 months, with the resurgence of the problem and in view of the phage and serotyping results, this view was revised. It has been suggested that in common source outbreaks a combination of 0 serotyping and phage typing normally suffices to recognize the epidemic strain and distinguish the strains which are distinct from it.‘5,‘7 In this outbreak also, these two epidemiological markers helped to eliminate the isolate from the air conditioner as the ‘index’ strain and indicate that the isolates from the equipment were the possible source of infection. In the final analysis it was concluded that the source of infection was the dialyser unit and machine and that the organism was probably carried to the vascular access sites through the hands of personnel handling these. Although we did not isolate any Serratia sp. from the hands of the personnel during surveillance, other workers too have suggested that transmission of Serratia from brushes to infants and from infant to infant has occurred through the hands of medical and nursing personnel even though culture of hands was negative.4 Since the ‘in-use’ dilution of formalin was found to be bactericidal for these serratia isolates we concluded that certain parts of the dialyser were not being disinfected by the formalin. Common-source outbreaks of sepsis due to Serr&ia and other Gram-negative bacteria, due to the persistance and multiplication of the organism in items of equipment, have also been cited by Farmer.i2 Pigmented strains of S. marcescens are postulated to be less virulent than non-pigmented strains. ’ The current outbreak was not associated with frequent severe infection probably because of the benign nature of the pigmented infecting strain. We would like to express our gratitude to Dr T. L. Pitt of the Division Central Public Health Laboratory, Colindale, London for performing typing of the serratia isolates.
of Hospital serotyping
Infection, and phage
References 1. Wilfert JN, Barrett FF, Kass EH. Bacteremia 1968; 279: 286-289.
due to Serratia
marcescens. N EngZJ Med
Serratia
marcescens
outbreak
61
2. Smith PT. Brookfield DSK. Shaw A. Gary I. An outbreak of Serratia marcescens in a neonatal-unit. Lancet 1984; ‘1: 15 l-l 53. - 3. Anaenostakis DT. Fitsialos C. Koutsia 1. Messaritakis. Matsaniotis N. A nursers outbreak of Serratia marcescen; infection: evidence of a single source of contamination. Am J Dis Child 1981; 135: 413414. 4. Wilhelmi I, Bernaldo de Quiros JCL, Romero-Vivas J, Duarte J, Rojo E, Bouza E. Epidemic outbreak of Serratia marcescens infection in a cardiac surgery unit. J Clin Microbial 1987; 25: 1298-1300. 5. Maki DG, Hennekens CG, Phillips CW, Shaw WV, Bennett JV. Nosocomial urinary tract infection with Serratia marcescens: an epidemiological study.g Infect Dis 1973; 128: 579-587. 6. Lowbury EJL, Ayliffe GAJ, Geddes AM, Williams JD. Control of Hospital Infection-A Practical Handbook, 2nd ed. London: Chapman and Hall 1981; 308-313. 7. Easterling RE. Mechanical aspects of dialysis including dialysate delivery systems and water for dialysate. In: Nissenson AR, Fine RN, Gentile DE, Eds. Clinical Dialysis. Connecticut: Appleton-Century-Crofts 1984; 55-61. 8. Cruickshank R, Duguid JP, Marmion BP, Swain RHA. Medical Microbiology Vol.II, 12th ed. London: Churchill Livingstone 1975; 191. 9. Finegold SM, Baron EJ. In Bailey and Scott’s Diagnostic Microbiology, 7th edn. St. Louis, USA: CM Mosby Company 1986; 379-383. 10. Barry LA, Thornsberry C. Susceptibility tests: diffusion test procedures. In: Lennete EH, Balows A, Hausler Jr WJ, Shadomy HJ, Eds. Manual of Clinical Microbiology, 4th edn. Washington DC: American Society for Microbiology 1985; 978-987. 11. Lewis DA, Hawkey PM, Watts JA et al. Infection with netilmicin resistant Serratia marcescens in a special baby care unit. Br Med J 1983; 287: 1701-1705. 12. Farmer JJ, Davis BR, Hickman FW et al. Detection of serratia outbreaks in hospital. Lancet 1976; 2: 455-459. 13. Whitby JL, Blair JN, Rampling A. Cross infection with Serratia marcescens in an intensive therapy unit. Lancet 1972; 2: 127-129. 14. Sanders Jr, CV, Luby JP, Johanson Jr WG, Barnett JA, Sanford JP. Serratia marcescens infections from inhalation therapy medications: nosocomial outbreak. Ann Intern Med 1970; 73: 15-21. 15. Pitt, TL. State of the art: typing of Serratia marcescens. J Hosp Infect 1982; 3: 9-14. 16. McGeer, A,, Low DE, Penner J, Ng J, Goldman C, Simor AE. Use of molecular typing to study epidemiology of Serratia marcescens. J Clin Microbial 1990; 28: 55-58. 17. Sifuentes-Osornio J, Ruiz-Palacios GM, Groschel DHM. Analysis of epidemiologic markers of nosocomial Serratia marcescens isolates with special reference to the Grimont biotyping system. J Clin Microbial 1986; 23: 23&234. “,
_I
of Hospital
Infection
(1991)
Epidemiological marcescens
study of an outbreak in a haemodialysis
P. U. Krishnan, Departments
18, 57-61
B. Pereira*
of Serratia unit
and R. Macaden
of Microbiology and *i?ephrology, St. John’s and Hospital, Bangalore 560034, India Accepted for publication
14 January
Medical
College
1991
Summary:
Twenty-three isolates of Serratia mamescens were isolated over a lo-month period from the blood and arteriovenous shunt sites of patients undergoing haemodialysis in an artificial kidney unit. Surveillance measures performed-on the equipment, sterile materials! environment and personnel of this unit vielded Serratia from the air condmoner and one of the dialvsis units. The isolates from the patients and dialyser unit were pigmented &d had an identical biochemical profile, antibiograms, phage typing pattern and 0 serotype. The isolate from the air conditioner, though of the same biotype, had a different phage and serotype. It was concluded that the dialyser was the ‘common source’ and that the organism was persisting in the machine in spite of recommended sterilization procedures being implemented. Keywords:
Sevratiu
marcescens; haemodialysis
unit;
epidemiological
survey.
Introduction Hospital-acquired infections by Serratia marccscens have been well documented in the immunocompromised host.‘,’ Over the past two decades, outbreaks of septicaemia, meningitis and wound infections have been increasingly reported, especially in patients subjected to invasive diagnostic and therapeutic procedures. 3*4The morbidity and mortality rates in these infections are generally high since most patients have serious underlying medical problems and the infecting strains are usually multiply drug resistant.4 Epidemiological marker studies have been done in an attempt to trace the source of infection in some of these outbreaks.4,5 Various types of outbreaks occur in the Haemodialysis Unit and bacterial infections-especially with hospital-acquired pathogens-are recognized to be the greatest hazard.6 The following report details the clinical and bacteriological features of an outbreak of serratia septicaemia in a Haemodialysis Unit over a lo-month period.
Correspondence 019s4701/91/050057+05
to: Dr P. U. Krishnan to3.00/0
0 1991 The Hospital
57
Infection
Society
58
P. U. Krishnan Materials
et al.
and methods
St. John’s Medical College Hospital is a 750-bed teaching hospital in South India. It has an artificial kidney unit which carries out about 150 haemodialysis procedures per month. Routine bacteriological survey of the dialysis fluid, re-use tubes and water treatment system is carried out once a month according to standard recommendations.7 The dialyser is re-used two or three times, after being disinfected by recommended techniques.6 Three cases of S. marcescens septicaemia were observed over a period of 2 weeks and these were brought to the notice of the Infection Control Committee. Surveillance measures were undertaken and samples were taken for culture from the water treatment system, haemodialysis fluid concentrate, re-use tubes, all equipment and furniture in the dialysis room and the air conditioner vents. The unit was temporarily closed for thorough disinfection and fumigation. Subsequently, on re-opening of the unit, three further cases of serratia infection either of blood and/or the arteriovenous shunt site occurred in the same unit over a period of 3 months, after which there were no further cases for 2 months. After this period and over the following 4 months, 15 sporadic cases of serratia septicaemia occurred in the dialysis unit. A second surveillance similar to the first was done. In addition, to test the bactericidal effect of 2% formalin (used for dialyser disinfection), a broth culture of the serratia isolate was kept in contact with this solution for 48 h at 37°C and then subcultured. Nail clippings and finger-web swabs were taken from Dialysis Unit personnel to detect carriers. The case records of all these patients were reviewed and analysed. All culture techniques, bacterial identification and antibiotic susceptibility tests were done according to recommended procedures.‘,” All the serratia isolates were sent to the Public Health Laboratory Service, London with a request to Dr T. L. Pitt for epidemiological marker studies. Results
In all, 26 isolates of S. maxescens were obtained, 21 from blood, two from vascular access sites, one from the environment (air conditioner vent), one from the dialysing machine No.1 and one from the contents of the dialyser kept ready for re-use. None of the hospital personnel were found to be carriers. The 2% formalin was found to be bactericidal for these serratia isolates. All the strains were pigmented, producing the typical cherry-red pigment prodigiosin. The biochemical profile of all strains, including the environmental one and those from the equipment, was identical. All isolates were susceptible to streptomycin (10 pg), choramphenicol (30 pg), kanamycin (30 pg), gentamicin (10 pg), cotrimoxazole (2.5 pg) and resistant to tetracycline (3Opg) and ampicillin (1Opg).
Serratia
marcescens
59
outbreak
All except the isolate from the air conditioner belonged to the 06 serological group (Table I) and had the same phage typing pattern isolate belonged to the 08 group and was (1/3/4/7/9/l 1). Th e air conditioner of a different phage type (l/3/5/7) All the patients had end stage renal failure and were undergoing dialysis in this Unit. In two of them Serratia was also isolated from the vascular access sites. Blood cultures were done in the first two cases when these patients presented with low-grade fever without localizing signs following haemodialysis. As the outbreak became established, blood cultures were done as a screening measure. S. marcescens was isolated from the blood cultures of 21 patients. In addition, swabs from the arteriovenous shunts of two of these patients yielded Serratia. Except for one patient who died of a massive cerebral infarction while on treatment, and another who developed bacterial endocarditis with the same organism, all the others had an uneventful recovery either on oral or parenteral antibiotic therapy. Discussion The incidence of S. marcescens infection in our hospital has never exceeded five cases per year over the last five years, and even among these, episodes of significant bacteraemia have been rare. This sudden and clustered increase in the isolation of Serratia from blood was a cause of alarm and various measures were taken to curb this outbreak. Fortunately, unlike previously described outbreaks, i,ll this strain was susceptible to commonly used antibiotics. Farmer et aZ.‘* have repor t ed that the initial resistance may usually be R-factor mediated and often includes resistance to the more useful, less toxic drugs such as tetracyclines, sulphonamides and ampicillin. Table
1. Analysis
Source
Blood
of Serratia
No. of isolates
culture
marcescens
isolates from patients
Antibiogram*
and environment
Serotype
Susceptibility pattern
Resistance pattern
Phage type
21
S,C,K,G,Cot
T,A
06
l/3/4/7/9/1
1
Vascular access site cultures
2
S,C,K,G,Cot
T,A
06
l/3/4/7/9/1
1
Air conditioner vent
1
S,C,K,G,Cot
T,A
08
l/3/5/7
Dialysing machine
1
S,C,K,G,Cot
T,A
06
l/3/4/7/9/1
1
1
S,C,K,G,Cot
T,A
06
l/3/4/7/9/1
1
Contents of dialyser
No. 1
*S, Streptomycin; C, chloramphenicol; tetracycline; A, ampicillin.
K, kanamycin;
G, gentamicin;
Cot,
cotrimoxazole;
T,
60
P. U. Krishnan
et al.
If these organisms are not eliminated, a vicious circle of increasing drug resistance will continue until a sub-strain resistant to all antibiotics emerges. In our hospital, however, even though the outbreak has continued for 10 months, multiply drug resistant strains have not emerged as yet. The source of infection has in previous outbreaks been traced to gastrointestinal carriage, especially in infants,” contaminated handwashing brushes,’ shaving brushes, l3 barber’s hands4 and inhalation therapy medications.14 However, it has been suggested that environmental reservoirs are insignificant as a source. l&l6 Initially, the air-conditioner was implicated as a common source because the isolate from that site had the same biochemical properties and antibiotic sensitivity pattern as the clinical isolates. Also, with the advent of the monsoon and the decreased use of the air-conditioner, the outbreak subsided. However, after a lapse of 2 months, with the resurgence of the problem and in view of the phage and serotyping results, this view was revised. It has been suggested that in common source outbreaks a combination of 0 serotyping and phage typing normally suffices to recognize the epidemic strain and distinguish the strains which are distinct from it.‘5,‘7 In this outbreak also, these two epidemiological markers helped to eliminate the isolate from the air conditioner as the ‘index’ strain and indicate that the isolates from the equipment were the possible source of infection. In the final analysis it was concluded that the source of infection was the dialyser unit and machine and that the organism was probably carried to the vascular access sites through the hands of personnel handling these. Although we did not isolate any Serratia sp. from the hands of the personnel during surveillance, other workers too have suggested that transmission of Serratia from brushes to infants and from infant to infant has occurred through the hands of medical and nursing personnel even though culture of hands was negative.4 Since the ‘in-use’ dilution of formalin was found to be bactericidal for these serratia isolates we concluded that certain parts of the dialyser were not being disinfected by the formalin. Common-source outbreaks of sepsis due to Serr&ia and other Gram-negative bacteria, due to the persistance and multiplication of the organism in items of equipment, have also been cited by Farmer.i2 Pigmented strains of S. marcescens are postulated to be less virulent than non-pigmented strains. ’ The current outbreak was not associated with frequent severe infection probably because of the benign nature of the pigmented infecting strain. We would like to express our gratitude to Dr T. L. Pitt of the Division Central Public Health Laboratory, Colindale, London for performing typing of the serratia isolates.
of Hospital serotyping
Infection, and phage
References 1. Wilfert JN, Barrett FF, Kass EH. Bacteremia 1968; 279: 286-289.
due to Serratia
marcescens. N EngZJ Med
Serratia
marcescens
outbreak
61
2. Smith PT. Brookfield DSK. Shaw A. Gary I. An outbreak of Serratia marcescens in a neonatal-unit. Lancet 1984; ‘1: 15 l-l 53. - 3. Anaenostakis DT. Fitsialos C. Koutsia 1. Messaritakis. Matsaniotis N. A nursers outbreak of Serratia marcescen; infection: evidence of a single source of contamination. Am J Dis Child 1981; 135: 413414. 4. Wilhelmi I, Bernaldo de Quiros JCL, Romero-Vivas J, Duarte J, Rojo E, Bouza E. Epidemic outbreak of Serratia marcescens infection in a cardiac surgery unit. J Clin Microbial 1987; 25: 1298-1300. 5. Maki DG, Hennekens CG, Phillips CW, Shaw WV, Bennett JV. Nosocomial urinary tract infection with Serratia marcescens: an epidemiological study.g Infect Dis 1973; 128: 579-587. 6. Lowbury EJL, Ayliffe GAJ, Geddes AM, Williams JD. Control of Hospital Infection-A Practical Handbook, 2nd ed. London: Chapman and Hall 1981; 308-313. 7. Easterling RE. Mechanical aspects of dialysis including dialysate delivery systems and water for dialysate. In: Nissenson AR, Fine RN, Gentile DE, Eds. Clinical Dialysis. Connecticut: Appleton-Century-Crofts 1984; 55-61. 8. Cruickshank R, Duguid JP, Marmion BP, Swain RHA. Medical Microbiology Vol.II, 12th ed. London: Churchill Livingstone 1975; 191. 9. Finegold SM, Baron EJ. In Bailey and Scott’s Diagnostic Microbiology, 7th edn. St. Louis, USA: CM Mosby Company 1986; 379-383. 10. Barry LA, Thornsberry C. Susceptibility tests: diffusion test procedures. In: Lennete EH, Balows A, Hausler Jr WJ, Shadomy HJ, Eds. Manual of Clinical Microbiology, 4th edn. Washington DC: American Society for Microbiology 1985; 978-987. 11. Lewis DA, Hawkey PM, Watts JA et al. Infection with netilmicin resistant Serratia marcescens in a special baby care unit. Br Med J 1983; 287: 1701-1705. 12. Farmer JJ, Davis BR, Hickman FW et al. Detection of serratia outbreaks in hospital. Lancet 1976; 2: 455-459. 13. Whitby JL, Blair JN, Rampling A. Cross infection with Serratia marcescens in an intensive therapy unit. Lancet 1972; 2: 127-129. 14. Sanders Jr, CV, Luby JP, Johanson Jr WG, Barnett JA, Sanford JP. Serratia marcescens infections from inhalation therapy medications: nosocomial outbreak. Ann Intern Med 1970; 73: 15-21. 15. Pitt, TL. State of the art: typing of Serratia marcescens. J Hosp Infect 1982; 3: 9-14. 16. McGeer, A,, Low DE, Penner J, Ng J, Goldman C, Simor AE. Use of molecular typing to study epidemiology of Serratia marcescens. J Clin Microbial 1990; 28: 55-58. 17. Sifuentes-Osornio J, Ruiz-Palacios GM, Groschel DHM. Analysis of epidemiologic markers of nosocomial Serratia marcescens isolates with special reference to the Grimont biotyping system. J Clin Microbial 1986; 23: 23&234. “,
_I
