=methicillin resistant staphylococcus aureus.  Cf methicillin sensitive staph, MSSA.

Methicillin resistance equates with flucloxacillin resistance. mecA is the methicillin resistance gene, which codes for a low affinity PBP (penicillin binding protein) – ie penicillin can’t bind easily. The gene has probably crossed from coag neg staph on at least 5 occasions to create MRSA strains. As with MSSA, different strains exist, carrying a range of different pathogenic genes.

Traditionally MRSA was found in institutions and the elderly, but now can be seen frequently in the young and healthy, causing the same infections that MSSA causes eg skin/soft tissue. It can also be responsible for rarer, more severe diseases eg necrotizing fasciitis. The US Center for Disease Control details criteria for distinguishing hospital acquired and community acquired MRSA infections – community acquired strains are typically SCCmec type IV (this is the cassette that contains mecA), which are sensitive to most non-beta lactam antibiotics, but on the other hand is associated with Panton Valentine Leucocidin (PVL, a cytotoxin associated with necrotizing disease). But again, this distinction is becoming less clear with strains associated with community acquired infection becoming more frequent in hospital acquired cases, and having variable levels of non-beta lactam antibiotic resistance.

In itself, antibiotic resistance may not translate to increased virulence and pathogenicity – it may just make it harder to treat. Studies have shown that after correcting for other factors eg age and co-morbidity, mortality is not significantly different. However, one important factor is use of inappropriate antibiotics, which of course is more likely with MRSA. Furthermore, in the US many MRSA outbreaks are caused by the USA300 clone, which carries a number of genes (in common with Methicillin sensitive staphylococcus aureus) eg PVL, ACME which are associated with enhanced pathogenicity.

PVL seen in 50% of symptomatic (skin) community MRSA in the US. Now also being reported in hospital acquired MRSA. Prevalence much lower in Europe, but risk of spread. Geographical areas tend to have their own clones (eg type 80 in Europe), with occasional pandemic.


  • MRSA has been shown to survive on sterile packaging for at least 6 months.Journal of hospital infection 2001;49(4):255-61.
  • Basic simple infection control like hand washing works.
  • MRSA prevalence in hospitals is associated with macrolide and 3rd generation cephalosporin use.Clinical microbiology and infection 2007;13(3):269-76.
  • Alcohol hand rub reduces its transmission and hospitals which have introduced a policy of using this between patient contacts reduce their MRSA rate.
  • Isolation and screening work but may be impracticable in emergency admissions or in hospitals with near 100% occupancy.

In the UK, most MRSA are resistant to quinolones and macrolides. Even if sensitive to quinolones, treatment is not recommended as resistance evolves rapidly. Most MRSAs are sensitive to tetracyclines (but not for use under 12yrs), rifampicin, co-trimoxazole, and linezolid, all of which can be given orally. Clindamycin can be effective but beware inducible resistance – see below. No evidence for trimethoprim alone (cf septrin); use in combination with rifampicin? Probably best not to give rifampicin or fusidic acid monotherapy anyway as resistance frequently induced.  See antibiotic classes.

In MRSA skin infections (cellulitis/abscesses), most will get better anyway, esp after I&D, but using the wrong antibiotic increases risk of treatment failure by odds ratio of 2.80 (87% success cf 95% of patients who received an active antibiotic). Topical agents will induce resistance if used for high load infections eg wounds, catheter sites so should be combined with systemic therapy.  About 12% are resistant to topical mupirocin.

For intravenous therapy, gentamicin, vancomycin and teicoplanin are effective, although vancomycin resistance has been described since 2002. Teicoplanin levels can be unpredictable, and treatment failure associated with low levels has been seen; checking levels would make sense (aiming for trough of at least 10, 20 in endocarditis) but is rarely done. In line infections, vancomycin or linezolid should be used if the infection is severe; if milder, then removal of the line and oral therapy may be sufficient.

In bone infection, linezolid is good but should be given for a maximum of 28 days. Fusidic acid and Rifampicin are good adjuncts (rifampicin has in vitro activity against biofilms). Clindamycin and co-trimoxazole have also been used for bone infections. Early surgery (eg within 2/7 of onset of symptoms) is important esp where a prosthesis is present.

Necrotising pneumonia with MRSA post-influenza has mortality up to 75%.

For bacteraemia/endocarditis, vancomycin is the drug of choice, as treatment failures have been described with teicoplanin. Rifampicin or fusidic acid can be considered as adjuncts; there is no evidence for adding an aminoglycoside. A minimum of 14 days treatment is required (although oral treatment may be appropriate for maintenance) but should be extended if vegetations seen on trans-oesophageal echo. Infected pacemakers should be considered the same as orthopaedic prostheses.

Clindamycin resistance is sometimes only 1 mutation away from erythromycin resistance. If the bug is erythromycin sensitive, then there is no issue, and clindamycin is a good choice (and can be given orally). On the other hand, if erythromycin resistance is seen, then the D test should be done: if a D-shaped zone appears around the clindamycin disk when an erythromycin disk is placed nearby, then you have induced resistance and clindamycin should be avoided. The erm gene is responsible for erythromycin-inducible resistance; the mrsA gene also confers resistance to erythromycin but does not affect clindamycin.


Once colonized, about 40% of patients develop persistence – commoner where skin breaks present. Vancomycin does not clear nose, throat or gut.

Eradication of S. aureus nasal colonization eg with 72 hour mupirocin has been successful. However, recolonization usually occurs within a relatively short time, and the Cochrane review did not find much evidence in favour. Use of mupirocin to prevent infection in endemic settings eg dialysis centers have shown conflicting results although metanalysis suggests benefit (but fear of mupirocin resistance). Neomycin is even less effective, but is an alternative where mupirocin resistance is seen.

Combined treatment seems sensible, and recent RCT of 2% chlorhexidine gluconate washes, 2% mupirocin ointment intranasally, oral rifampin and doxycycline for 7 days vs no treatment confirms. At 3 months of follow-up, 74% cf 32% had cleared. Still significant benefit at 8 months (54% of those treated culture negative). On multivariable analysis, having a mupirocin-resistant isolate increased the risk of treatment failure nearly 10 fold. Mupirocin resistance emerged in only 5% of follow-up isolates. Clin Infect Dis. 2007 Jan 15;44(2):178-85.

Other control measures include a combination of active surveillance cultures of high risk patients, improved health care worker hand hygiene, consistent use of contact precautions for colonized/infected patients, and directed treatment of health care workers implicated in transmission. PIDJ January 2005 pp 79-80

Screening patients seems to reduce hospital acquired infection in the Netherlands, but not in Switzerland. UK guidelines say do for high risk only eg previous carriers, transfers, ICU. Standard infection control procedures alone seem to have worked in UK although what do you compare with? Rapid test (PCR) did not help (in crossover trial) except in reducing inappropriate isolation. 3-4x as expensive.

[UK guidelines, J antimicro chemo 2006 PMID 16507559]