Utility of spa typing for investigating the local epidemiology of MRSA on a UK intensive care ward

Journal of Hospital Infection (2009) 71, 29e35 Available online at www.sciencedirect.com www.elsevierhealth.com/journals/jhin Utility of spa typing for investigating the local epidemiology of MRSA on a UK intensive care ward S. Khandavilli a, P. Wilson b, B. Cookson c, J. Cepeda c, G. Bellingan a, J. Brown a, * a Centre for Respiratory Research, Department of Medicine, Royal Free and University College Medical School, Rayne Institute, London, UK b Department of Clinical Microbiology, University College London Hospitals, London, UK c Specialist and Reference Microbiology Division, Health Protection Agency, London, UK Received 18 April 2008; accepted 12 September 2008 Available online 17 November 2008 KEYWORDS Multilocus sequence typing; Meticillinresistant Staphylococcus aureus; spa typing Summary In the UK, meticillin-resistant Staphylococcus aureus (MRSA) is frequently endemic on intensive care units (ICUs), yet our understanding of the local epidemiology of MRSA within the ICU is poor and the best methods for preventing MRSA acquisition remain controversial. Newer molecular typing methods may aid epidemiological investigation of local MRSA strains. We applied Staphylococcal Protein A (spa) typing to MRSA strains collected from patients in a UK ICU. spa typing allowed better discrimination than multilocus sequence typing (MLST) but 73% of strains were either spa type t032 or t018 (associated with the prevalent UK MRSA strains, EMRSA-15 and EMRSA-16). MRSA infections were preceded by MRSA colonisation in 72% of patients, and in 88% of these, both commensal and diseasecausing strains had identical MLST and spa types. spa typing helped elucidate the transmission of MRSA between patients for 19 strains with unusual spa types, although the high incidence of EMRSA-15 and -16 types t032 and t018 prevented its use for the majority of strains. Surprisingly, only four (9%) of 45 new MRSA isolates occurring within 28 days of isolation of an unusual spa type could have been due to cross-contamination. These results suggest that prompt transmission of MRSA between patients is rare in our ICU, at least for those strains with unusual spa types. ª 2008 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved. * Corresponding author. Address: Centre for Respiratory Research, Department of Medicine, Royal Free and University College Medical School, Rayne Institute, 5 University Street, London WC1E 6JJ, UK. Tel.: þ44 20 7679 6008; fax: þ44 20 7679 6973. E-mail address: jeremy.brown@ucl.ac.uk 0195-6701/$ - see front matter ª 2008 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jhin.2008.09.011

30 S. Khandavilli et al. Introduction Meticillin-resistant Staphylococcus aureus (MRSA) has attracted attention as a nosocomial pathogen causing considerable morbidity and mortality on a global scale. 1 It causes one-fifth of nosocomial infections in the UK and up to one in six patients in UK intensive care unit (ICU) wards are colonised or infected with MRSA. These strains are usually EMRSA-15 and -16, the most prevalent epidemic MRSA strains in the UK. 2e5 MRSA is thought to be transmitted from patient to patient via the hospital staff or environment but there is surprisingly little published detailed data on the spread of particular MRSA strains from patient to patient. How readily MRSA are disseminated from an index case to other patients on an ICU ward is not known, and at present the most effective method to prevent MRSA infection is controversial. 6 Isolation of patients, when part of a larger package of control measures, is associated with reduced transmission of MRSA, but isolation of MRSA-positive patients in single rooms is not standard policy for 24% of English ICUs. 3,7 Furthermore, a prospective study from our centre has shown that isolating or cohort nursing MRSA patients in an ICU with endemic MRSA infection does not reduce cross-infections, although hand hygiene practices were less than optimal. 8 Studies on local transmission requires good clinical data, local surveillance of MRSA, and a method of typing different MRSA strains with adequate discrimination. Several typing methods have been developed to aid investigation of MRSA epidemiology. Discrimination between strains by phage typing is very strain dependent and lacks reproducibility, making the identification of clonal relationships between strains difficult without confirmation by molecular techniques. 9 Pulsedfield gel electrophoresis (PFGE) is an effective method for investigating the local epidemiology of MRSA strains, but is labour intensive and, in common with phage typing, the results are difficult to standardise between laboratories. 9 Multilocus sequence typing (MLST), a molecular technique based on DNA sequencing of housekeeping genes, allows more precise identification of a particular strain and ready comparison of results between different laboratories. 10 MLST has provided valuable insights into the national and international epidemiology of MRSA but lacks the discriminatory power for investigating local epidemiology when there is a high prevalence of epidemic strains such as EMRSA-15 or -16. Staphylococcal Protein A (spa) typing requires the sequencing of only one locus, the short sequence repeat region of the gene that encodes Protein A, and has a greater discriminatory power than MLST. spa typing can define the local epidemiology of MRSA infection in various settings, but whether it has the discriminatory power to investigate cross-infection in a hospital environment dominated by EMRSA-15 and -16 has not been assessed. 11e17 In addition, although a substantial proportion of all Staphylococcus aureus strains isolated from the blood have identical PFGE patterns to nasal isolates from the same patient, we are not aware of studies specifically investigating whether MRSA infections are caused by strains already colonising the patient or by strains acquired from exogenous sources. In the ICU, exogenous acquisition of MRSA may be more prevalent than in a general hospital population due to high levels of environmental contamination and staff carriage. 18,19 This study assesses whether spa typing can be used to investigate the epidemiology of MRSA within a UK ICU, using microbiological and clinical data obtained during a clinical trial of different methods of controlling MRSA infection among ICU patients. 8 In addition, we investigated whether the colonising strain is the same as the strain causing infection for patients with invasive MRSA infection, as well as the degree of spread between patients. Methods Bacterial isolates and culture The MRSA strains used for this study were isolated from patients during a 12 month clinical study on the effectiveness of case isolation for controlling MRSA infection of patients in a London-based ICU. 8 The infection control procedures used during this study included disposable aprons, gloves for patient contact, and for half the study period only, isolation of MRSA-positive patients in single rooms or in an open bay of other patients who were MRSA positive. 8 Strains were stored on beads at e70 C (Microbank system, Pro Lab Diagnostics, Neston, UK). Patients were screened (nasal/perineal swabs, NPS) for MRSA within 24 h of admission to the ICU, weekly thereafter and at discharge. Clinical samples (wound swabs and blood cultures) were obtained when clinically indicated. During the trial period, 746 MRSA isolates were obtained from our hospital, phage typing of which suggested 219 (29%) were EMRSA-15 and 271 (36%) were EMRSA-16. 8 For this study, a total of 115 of the

spa typing of UK intensive care MRSA isolates 31 746 MRSA isolates underwent MLST and spa typing. These 115 strains consisted of the following distinct groups: (i) 45 isolates randomly selected to represent each of the phage types present within the NPS collection, with over-representation of the common phage types; (ii) all 43 of invasive isolates, defined as MRSA strains cultured from wound swabs (25 patients) or the blood (18 patients); and (iii) an additional 27 commensal NPS isolates isolated from 43 patients with invasive isolates. As well as the above 115 isolates, all MRSA isolates acquired by ICU patients within 28 days of isolation of an MRSA strain with an unusual spa type from another ICU patient underwent spa typing alone (totalling 15 isolates). MRSA isolates were cultured in Brain Heart Infusion (BHI) broth (Sigma, Dorset, UK) at 37 C and 220 rotations per minute on a rotary shaker overnight. DNA extraction, spa and MLST typing Genomic DNA was extracted from overnight grown cultures using Wizard Genomic DNA extraction kit (Promega, Southampton,UK) according to the manufacturer s instructions except that 30 mg/l lysostaphin (Sigma) was added at the lysis step. spa typing was performed according to the method of Shopsin et al. using the primers SpaF: AGACGATCCTTC GGTGA and SpaR: CAGCAGTAGTGCCGTTTG. 20 Isolates were assigned to particular spa types using the spa typing website (http://www.spaserver. ridom.de) developed by Ridom GmbH and curated by SeqNet.org (http://www.seqnet.org). 16 MLST typing was performed according to the protocol described by Enright et al. 10 and isolates assigned a sequence type (ST) according to the MLST website http://www.mlst.net. PCR products were purified by precipitation using 20% polyethylene glycol in 2.5 M NaCl and both strands were sequenced using Big dye fluorescent terminator V1.1 by an ABI prism 3700 DNA sequencer (Applied Biosytems, Cheshire, UK). Results Spa and MLST typing of NPS isolates The utility of spa typing at discriminating between MRSA strains isolated from patients in the UK was investigated using 45 MRSA NPS isolates representing all the phage types in our strain collection, with common phage types represented by several strains. These isolates were typed using spa and MLST and found to include 10 spa types, including one previously unreported type (t1302), and seven MLST types (Table I). The majority (39,87%) of the strains were divided by MLST into ST22 (representing EMRSA-15) and ST36 isolates (representing EMRSA-16). In addition, two of the other STs identified were single locus variants of ST22 or ST36 (ST217 and ST30 respectively), suggesting that these strains were also derived from EMRSA-15 or -16. 10 spa typing subdivided the EMRSA-15 ST22 strains into four different spa types (t022, t032, t1287 and t1302) and the EMRSA- 16 ST36 strains into two spa types (t012 and t018). However, the majority of the ST22 and ST36 strains were spa type t032 (71%) and t018 (92%) respectively, types that have previously been reported to represent EMRSA-15 and -16 (http://www.spaserver.ridom.de). Relationship between infecting and commensal MRSA strains To establish the relationship between MRSA colonisation and infection, we identified patients in whom MRSA was isolated by NPS within 14 days of the subsequent isolation of MRSA causing an infection (defined as MRSA isolated from a wound swab or blood culture). Of the total of 43 patients with infection due to MRSA during the study, 31 (72%) had a positive NPS for MRSA in the Table I Breakdown of results of phage, multilocus sequence typing (ST) and spa typing for 45 nasal/perineal swab meticillin-resistant Staphylococcus aureus isolates (numbers of each isolate type given in parentheses) Phage type ST spa types E15 (7) 22 (14) t032 (10) NT a (6) t022 (2) Dis b (1) t1287 (1) t1302 (1) E16 (11) 36 (25) t018 (23) 16V (3) t012 (2) 16D (3) NK c (3) B (1) NT a (1) 47 (1) Dis b (1) T (1) A (2) ST8 t190 ST111 t041 SMA ST111 t041 II ST188 t189 NK c (2) ST30 t1111 ST217 t1287 a NT, not typable; unreactive to the 27 phages used. b Dis, distinct; unrecognised phage pattern. c NK, not known: no clear results obtained.

32 S. Khandavilli et al. preceding two weeks (median seven days). Four of these NPS isolates were not available for study, leaving 27 matched NPS/invasive isolates which were typed using the spa and MLST methods. For 24 (88%) of these pairs, both the carriage and disease-causing MRSA strains had identical ST and spa types. Of the remaining three pairs, strains had the same ST but a different spa type for two pairs and both the ST and spa type were different for the other pair. Local spread of strains with unusual spa types Of the total of 115 strains described above which were characterised using spa and MLST typing (45 NPS isolates, 43 infection isolates and 27 NPS isolates from patients with invasive isolates), 24 (21%) were either EMRSA-15 or -16 on MLST typing but belonged to one of seven unusual spa types only present in a small number of patients (t012, 11; t1214, four; t1287, three; t022, three; t1302, two; and t253, one). These unusual spa types were rare enough to suggest that isolation of MRSA with the same unusual spa type from patients concurrently on the ICU could represent patient-to-patient spread. We therefore used the clinical details available for each patient recorded during the controlled trial, including the temporal relationship between admission to ICU and the acquisition of MRSA, to investigate whether identical unusual spa types were isolated from other patients present on the ICU concurrently or within 28 days. 8 There were 24 index cases of patients infected with EMRSA-15 or -16 of an unusual spa type, and we investigated all MRSA isolates from other ICU patients within 28 days of these index cases. For three index cases (all spa type t012) lack of clinical data prevented tracking of MRSA acquisition cases, and for a further two (both t012) no patients acquired MRSA within 28 days of the index case. For the remaining 19 index cases, within 28 days of isolation of an unusual spa type NPS, wound swabs, sputum or blood cultures were positive for MRSA in a total of 45 previously MRSA-negative patients (excluding cases occurring in patients within 48 h of admission to the ICU as they were probably already colonised). There was a range of one to six new acquisition MRSA cases per index case, predominantly belonging to the EMRSA-16 and EMRSA-15 spa types t018 (24 cases) and t032 (nine cases) respectively with only 12 (27%) isolates having an unusual spa type (Table II). Furthermore, the unusual spa type in acquisition cases matched the unusual spa type isolated from the corresponding index case in only four (9%) cases (Table II). For all four cases the index case was the commonest unusual spa type, t012 (17% of EMRSA-16 isolates), and this spa type was isolated from either one or two acquisition cases from 12 to 23 days after isolation of spa type t012 from the index case. Discussion Nosocomial infection with MRSA is a major health issue in the UK, yet important questions on the epidemiology of MRSA carriage and infection remain unanswered. A variety of molecular typing methods have been devised to track the spread of S. aureus infection. 2 Phage typing and PFGE are highly discriminatory but poorly reproducible between laboratories, whereas the sequencing techniques MLST and spa typing are reproducible but lack discriminatory ability. 9 spa typing has better discrimination than MLST and is cheaper, requiring a single sequencing reaction. 21 We studied whether spa typing can assess the local epidemiology of MRSA in a UK ICU setting by typing 45 MRSA isolates from ICU patients chosen to represent the full range of phage types in our collection. The results show that spa typing has greater discriminatory power than MLST, but, despite deliberately investigating a large range of phage types, 73% of strains still belonged to two spa types, t032 and t018, representing EMRSA-15 and -16 respectively. Hence spa typing, like MLST, is unlikely to be a useful technique to follow the spread of MRSA strains between patients where there is a high prevalence of EMRSA-15 or -16 strains. Despite the predominance of spa types t018 and t032, the significant numbers of rarer spa types isolated allowed us to investigate whether patient-topatient transmission of MRSA was common. Of 45 new cases of MRSA occurring in the ICU after isolation of an EMRSA-15 or -16 strain with an unusual spa type, only four (9%) could have been due to transmission of the unusual spa type EMRSA-15 or -16. It is possible, though unlikely, that the unusual spa type strains are defective in their ability to cross-infect, and transmission of EMRSA strains between patients is limited to the two common spa types t018 and t032. Our data suggest that on an ICU with high levels of endemic MRSA, rapid transmission of MRSA between patients is relatively rare, provided the usual infection control practices are employed. 8 We followed practices recommended by the US Hospital Infection Control Practices Advisory Committee and the UK Working Party (although only a proportion of MRSA-positive patients were isolated). Alternative explanations as to why

spa typing of UK intensive care MRSA isolates 33 Table II Results of spa typing for acquisition cases of meticillin-resistant Staphylococcus aureus (MRSA) with positive samples within 28 days of an index case of MRSA with an unusual spa type Case no. spa type Index cases No. of days post-admission to ICU MRSA isolated MRSA source spa type No. of days post-index case MRSA isolated Acquisition cases No. of days post-admission to ICU MRSA isolated MRSA source 1 t022 1 NPS t032 7 16 NPS t1287 14 4 NPS 2 t022 1 NPS t018 11 25 Sputum 3 t022 1 NPS t032 6 5 NPS t032 8 3 NPS t018 28 9 NPS 4 t1214 15 NPS t1302 4 10 NPS t018 21 11 NPS t018 10 3 NPS 5 t1214 9 Wound t012 8 6 NPS 6 t1214 20 Blood t018 5 3 NPS 7 t1214 15 Wound t018 11 3 NPS t018 22 11 NPS t018 25 3 NPS 8 t1287 5 Blood t018 6 8 Blood t018 9 25 Sputum 9 t1287 33 Wound t018 1 3 NPS t032 2 3 NPS 10 t1287 4 NPS t1214 6 19 Sputum t012 3 6 Wound 11 t1302 3 NPS t018 16 8 NPS t018 18 5 Blood 12 t1302 10 NPS t012 22 25 Blood t018 17 11 NPS t018 6 3 NPS 13 t253 16 Wound t012 15 4 NPS t018 10 12 NPS 14 t012 6 NPS t018 13 5 NPS 15 t012 2 NPS t032 13 4 NPS t012 a 23 9 NPS t032 28 11 Sputum 16 t012 9 NPS t012 a 12 4 Blood t012 a 12 4 Sputum t018 7 3 Sputum 17 t012 16 NPS t012 a 15 4 NPS t018 10 12 NPS 18 t012 6 NPS t018 5 27 NPS t018 25 25 Sputum t018 22 8 NPS t1287 16 5 Blood t018 3 10 Sputum t018 8 34 NPS 19 t012 2 Blood t032 13 3 NPS t032 13 3 Blood t032 28 10 Sputum ICU, intensive care unit; NPS, nasal/perineal swab. a Acquisition cases with the same spa type as the corresponding index case. patients may acquire MRSA on the ICU include the unmasking of low levels of endogenous MRSA colonisation by antibiotic use, or acquisition of MRSA from an intermediate reservoir such as environmental sources or colonised staff rather than transmission from other patients. 22,23 Further research is required to clarify the source of MRSA infection for ICU patients to provide the data

34 S. Khandavilli et al. necessary for the development of effective prevention policies. We also showed that MRSA infection is preceded by colonisation with an MRSA strain that is genetically indistinguishable from the disease-causing isolate in at least 56% of patients. This may be an underestimate as NPS samples were not obtained on a daily basis and strains were not available for typing for four patients with MRSA infection. However, this proportion may also be slightly overestimated due to the relative lack of variation in spa and ST types within our population, allowing identical spa and MLST types to occur in NPS and infection isolates from the same patient by chance. The results are similar to those obtained from previous studies of S. aureus infection, and suggest that MRSA infection usually follows prior carriage rather than occurring from direct transmission during invasive procedures by staff or from the ICU environment. 18,19 These data support the view that prevention of colonisation of ICU patients with MRSA could reduce the frequency of MRSA infections. 24 To summarise, we used spa typing to investigate the epidemiology of MRSA infection on a UK ICU. Although spa typing has better discrimination than MLST, the high prevalence of EMRSA-15 and -16 infections prevents this method from being a useful epidemiological tool in this setting. Monitoring EMRSA-15 and -16 infections with unusual spa types suggests that rapid patient-to-patient spread of MRSA on the ICU is a relatively rare event, although these data need to be interpreted cautiously due to the limitations of the molecular typing methods used and the possibility that unusual strains may have properties distinct from that of commoner MRSA strains. In addition, strains causing infection are usually due to endogenous commensal strains rather than direct invasion by exogenous strains. These data will assist the design of effective prevention strategies against MRSA infection. Acknowledgements We thank the Staphylococcal Section of the Laboratory of Healthcare Associated Infection, Health Protection Agency for phage typing the isolates. Conflict of interest statement None. Funding sources This work was supported by a grant from the University College London Hospitals Charities. References 1. Ayliffe GA. The progressive intercontinental spread of methicillin-resistant Staphylococcus aureus. Clin Infect Dis 1997;24(Suppl. 1):S74eS79. 2. National Audit Office. The management and control of hospital acquired infection in acute NHS trusts in England: report by Controller and Auditor General. London: Stationery Office; 2000. Paragraphs 15e16. 3. Hails J, Kwaku F, Wilson AP, Bellingan G, Singer M. Large variation in MRSA policies, procedures and prevalence in English intensive care units: a questionnaire analysis. Intensive Care Med 2003;29:481e483. 4. Johnson AP, Aucken HM, Cavendish S, et al. Dominance of EMRSA-15 and -16 among MRSA causing nosocomial bacteraemia in the UK: analysis of isolates from the European Antimicrobial Resistance Surveillance System (EARSS). J Antimicrob Chemother 2001;48:143e144. 5. Moore PC, Lindsay JA. Molecular characterisation of the dominant UK methicillin-resistant Staphylococcus aureus strains, EMRSA-15 and EMRSA-16. J Med Microbiol 2002;51:516e521. 6. Barakate MS, Yang YX, Foo SH, et al. An epidemiological survey of methicillin-resistant Staphylococcus aureus in a tertiary referral hospital. J Hosp Infect 2000;44:19e26. 7. Cooper BS, Stone SP, Kibbler CC, et al. Isolation measures in the hospital management of methicillin resistant Staphylococcus aureus (MRSA): systematic review of the literature. Br Med J 2004;329:533. 8. Cepeda JA, Whitehouse T, Cooper B, et al. Isolation of patients in single rooms or cohorts to reduce spread of MRSA in intensive-care units: prospective two-centre study. Lancet 2005;365:295e304. 9. Weller TM. Methicillin-resistant Staphylococcus aureus typing methods: which should be the international standard? J Hosp Infect 2000;44:160e172. 10. Enright MC, Day NP, Davies CE, Peacock SJ, Spratt BG. Multilocus sequence typing for characterization of methicillin-resistant and methicillin-susceptible clones of Staphylococcus aureus. J Clin Microbiol 2000;38:1008e1015. 11. Heym B, Le Moal M, Armand-Lefevre L, Nicolas-Chanoine MH. Multilocus sequence typing (MLST) shows that the Iberian clone of methicillin-resistant Staphylococcus aureus has spread to France and acquired reduced susceptibility to teicoplanin. J Antimicrob Chemother 2002;50:323e329. 12. Enright MC, Robinson DA, Randle G, Feil EJ, Grundmann H, Spratt BG. The evolutionary history of methicillin-resistant Staphylococcus aureus (MRSA). Proc Natl Acad Sci USA 2002;99:7687e7692. 13. Frenay HM, Bunschoten AE, Schouls LM, et al. Molecular typing of methicillin-resistant Staphylococcus aureus on the basis of protein A gene polymorphism. Eur J Clin Microbiol Infect Dis 1996;15:60e64. 14. Oliveira DC, Crisostomo I, Santos-Sanches I, et al. Comparison of DNA sequencing of the Protein A gene polymorphic region with other molecular typing techniques for typing two epidemiologically diverse collections of methicillin-resistant Staphylococcus aureus. J Clin Microbiol 2001;39:574e580. 15. Moodley A, Stegger M, Bagcigil AF, et al. spa typing of methicillin-resistant Staphylococcus aureus isolated from domestic animals and veterinary staff in the UK and Ireland. J Antimicrob Chemother 2006;58:1118e1123. 16. Harmsen D, Claus H, Witte W, et al. Typing of methicillinresistant Staphylococcus aureus in a university hospital setting by using novel software for spa repeat determination and database management. J Clin Microbiol 2003;41: 5442e5448.

spa typing of UK intensive care MRSA isolates 35 17. Saiman L, Cronquist A, Wu F, et al. An outbreak of methicillin-resistant Staphylococcus aureus in a neonatal intensive care unit. Infect Control Hosp Epidemiol 2003;24: 317e321. 18. von Eiff C, Becker K, Machka K, Stammer H, Peters G. Nasal carriage as a source of Staphylococcus aureus bacteremia. Study group. N Engl J Med 2001;344:11e16. 19. Chang FY, Singh N, Gayowski T, Drenning SD, Wagener MM, Marino IR. Staphylococcus aureus nasal colonization and association with infections in liver transplant recipients. Transplantation 1998;65:1169e1172. 20. Shopsin B, Gomez M, Montgomery SO, et al. Evaluation of protein A gene polymorphic region DNA sequencing for typing of Staphylococcus aureus strains. J Clin Microbiol 1999; 37:3556e3563. 21. Wisplinghoff H, Ewertz B, Wisplinghoff S, et al. Molecular evolution of methicillin-resistant Staphylococcus aureus in the metropolitan area of Cologne, Germany, from 1984 to 1998. J Clin Microbiol 2005;43:5445e5451. 22. Lemmen SW, Hafner H, Zolldann D, Stanzel S, Lütticken R. Distribution of multi-resistant Gram-negative versus Grampositive bacteria in the hospital inanimate environment. J Hosp Infect 2004;56:191e197. 23. Wilson AP, Hayman S, Whitehouse T, et al. Importance of the environment for patient acquisition of methicillin-resistant Staphylococcus aureus in the intensive care unit: a baseline study. Crit Care Med 2007;35:2275e2279. 24. Keshtgar MR, Khalili A, Coen PG, et al. Impact of rapid molecular screening for meticillin-resistant Staphylococcus aureus in surgical wards. Br J Surg 2008;95:381e386.