Importance of Control Group Selection for Evaluating Antimicrobial Use as a Risk Factor for Methicillin-ResistantStaphylococcus AureusBacteremia

Fluoroquinolone use is a risk factor for methicillin-resistant Staphylococcus aureus acquisition in long-term care facilities: a nested case-case-control study

BACKGROUND

Methicillin-resistant Staphylococcus aureus (MRSA) nasal colonization is a well-established risk factor for subsequent infection and a key event in interindividual transmission. Some studies have showed an association between fluoroquinolones and MRSA colonization or infection. The present study was performed to identify specific risk factors for MRSA acquisition in long-term care facilities (LTCFs).

METHODS

A prospective cohort of patients naive for S. aureus colonization was established and followed (January 2008 through October 2010) in 4 French LTCFs. Nasal colonization status and potential risk factors were assessed weekly for 13 weeks after inclusion. Variables associated with S. aureus acquisition were identified in a nested-matched case-case-control study using conditional logistic regression models. Cases were patients who acquired MRSA (or methicillin-sensitive S. aureus [MSSA]). Patients whose nasal swab samples were always negative served as controls. Matching criteria were center, date of first nasal swab sample, and exposure time.

RESULTS

Among 451 included patients, 76 MRSA cases were matched to 207 controls and 112 MSSA cases to 208 controls. Multivariable analysis retained fluoroquinolones (odds ratio, 2.17; 95% confidence interval, 1.01-4.67), male sex (2.09; 1.10-3.98), and more intensive care at admission (3.24; 1.74-6.04) as significantly associated with MRSA acquisition, and body-washing assistance (2.85; 1.27-6.42) and use of a urination device (1.79; 1.01-3.18) as significantly associated with MSSA acquisition.

CONCLUSIONS

Our results suggest that fluoroquinolones are a risk factor for MRSA acquisition. Control measures to limit MRSA spread in LTCFs should also be based on optimization of fluoroquinolone use.

Risk factors for methicillin-resistant Staphylococcus aureus bacteraemia differ depending on the control group chosen

Methicillin-resistant Staphylococcus aureus (MRSA) bacteraemia cause significant morbidity and mortality in hospitalized patients. Using a nested case-control design, 204 MRSA bacteraemia cases were compared to 301 unmatched methicillin-susceptible Staphylococcus aureus (MSSA) bacteraemia controls and were matched 1:2 with non-infected controls. The independent risk factors for MRSA bacteraemia compared to MSSA bacteraemia were older age (P = 0·048), major organ transplant during current hospital stay (P = 0·016) and quinolone use (P = 0·016). Cases were more likely than non-infected controls to have renal failure (P = 0·003), cirrhosis (P = 0·013), and a central venous catheter (P = 0·003) after controlling for other risk factors. This large case-control study made it possible to assess risk factors for MRSA bacteraemia using two sets of controls and showed that risk factors differed greatly depending on the control group chosen. These results confirm the need for careful selection of appropriate control groups and the need to carefully adjust for underlying severity of illness.

Length of exposure to the hospital environment is more important than antibiotic exposure in healthcare associated infections by methicillin-resistant Staphylococcus aureus: a comparative study

OBJECTIVES

Both total antimicrobial use and specific antimicrobials have been implicated as risk factors for healthcare-associated methicillin-resistant Staphylococcus aureus (HCA-MRSA) infection. The aims of this study were: (I) to explore predictors of a new HCA-MRSA infection in comparison with a new healthcare-associated methicillin-sensitive Staphylococcus aureus (HCA-MSSA); (II) to thoroughly assess the role of recent antibiotic use qualitatively and quantitatively.

METHODS

The time-period for our study was from October 1997 through September 2001. Through applying strict criteria, we identified two groups of inpatients, one with a new HCA-MRSA infection and one with a new HCA-MSSA infection. We recorded demographic, clinical and antibiotic use-related data up to 30 days before the positive culture date.

RESULTS

We identified 127 and 70 patients for each group, respectively. Two logistic regression models were carried out to assess the role of antimicrobial use (qualitatively and quantitatively). In model I, duration of hospital stay, presence of chronic wounds, aminoglycoside and fluoroquinolone use retained statistical significance. In model II, duration of hospital stay and history of intubation during the last month stood out as the only significant predictors of a subsequent HCA-MRSA infection. No significant differences in outcome were noted.

CONCLUSIONS

The length of exposure to the hospital environment may be the best predictor of a new HCA-MRSA infection. Use of aminoglycosides and fluoroquinolones may also stand independently along with presence of chronic ulcers and surgical procedures. No independent association between quantitative antibiotic use and subsequent HCA-MRSA infection was documented.

Role of matching in case-control studies of antimicrobial resistance

Of 57 case-control studies of antimicrobial resistance, matching was used in 23 (40%). Matched variables differed substantially across studies. Of these 23 matched case-control studies, 12 (52%) justified the use of matching, and 9 (39%) noted the strengths or limitations of this approach. Analysis that accounted for matching was performed in only 52% of the case-control studies.

Matching criteria in case-control studies in the field of antimicrobial resistance

OBJECTIVE

We evaluated the available evidence from case-control studies in the field of antimicrobial resistance to identify the degree that matching was performed and the criteria used to do so.

METHODS

We performed a systematic search of the PubMed database (articles archived by 08/2006) to identify relevant studies. Studies that used the individual matching technique were further analyzed.

RESULTS

115 case-control studies with a focus on antimicrobial resistance were identified: 28 regarding Acinetobacter baumannii, 25 regarding Pseudomonas aeruginosa, and 62 for other bacteria. Individual matching was performed in 32 (27.8%) out of the 115 studies. Age was the most frequently used matching criterion in 22 of 32 (69%) evaluated matched case-control studies, while sex was used in 11 (34%), presence of underlying illness in 8 (25%), site of infection in 5 (16%), and area of residence in 4 studies (12.5%). Other criteria were used in less than 10% of the studies. Analysis of data for matched pairs was performed in 18 of 32 (56.3%) studies that used a matched case-control design.

DISCUSSION

The available evidence from the analyzed data from case-control studies in the field of antimicrobial resistance shows that individual matching is employed only in a proportion of such studies and only for a few characteristics, when it was used. Also, analysis for matched pairs was not used by almost a half of the relevant studies.

Antibiotic resistance: location, location, location

Antibiotic resistance surveys are published widely, citing percentage resistance rates, sometimes for vast transcontinental regions. Such data seem straightforward, but when one drills deeper, great complexity emerges. Rates for methicillin resistance among Staphylococcus aureus from bacteraemias vary from <1% to 50% among European countries, and vary greatly among both hospitals and hospital units. Methicillin-resistant S. aureus (MRSA) resistance rates are typically higher for tertiary-care hospitals and intensive care units than in general hospitals and wards, and lowest in single specialist centres. The likelihood of resistance also varies according to patient characteristics: those patients from nursing homes and with underlying disease, recent antibiotic treatment and hospitalisation are more likely to harbour resistant pathogens. Percentage rates themselves also may be misleading; they may be high only because the denominator is small or inaccurate; i.e., resistance may be common but the pathogen rare. Measures of disease burden-cases per 1000 bed-days or per 10(5) individuals-overcome this deficiency but are harder to collect, influenced by case mix, and associated with other problems: how to count part days or infections acquired elsewhere; most important, are all cases captured? National or international resistance statistics may illustrate trends and provide benchmarks, but for patient management, good local data are essential. Which units are most affected? Are the resistant infections locally acquired or imported with transferred patients? Are the resistant isolates clonal, indicating cross-infection, or diverse, indicating repeated selection or reflecting antibiotic policy? Unless these aspects of infection are considered, interventions to reduce resistance may be misdirected.