Konsequenzen für Bevölkerung und Gesundheitssysteme Stephan Harbarth Infection Control Program University of Geneva Hospitals
Outline Introduction What data sources are available? AMR-associated outcomes Current data about impact? Excess costs of MRSA? Methodological challenges: How to determine clinical and public health impact Choice of methods? Confounding?
How to quantify the public health burden of a disease Population-based excess mortality Direct and indirect costs Lost DALYs (disability-adjusted life years) & QALYs (quality-adjusted life years)
10 leading causes of lost DALYs (15-44 y) in 1990 for developed countries Females DALYs ('000) Males DALYs ('000) All causes 24,674 All causes 36,943 Unipolar major 4,910 Alcohol use 4,677 depression Schizophrenia 1,450 Road traffic 4,167 accidents Road traffic 1,137 Unipolar major 2,664 accidents depression Bipolar disorder 1,106 Self-inflicted 2,072 injuries Obsessivecompulsive 933 Schizophrenia 1,578 disorders Alcohol use 801 Drug use 1,404 Osteoarthritis 783 Violence 1,196 Chlamydia 599 Ischaemic heart 1,160 disease Self-inflicted injuries 569 Bipolar disorder 1,135 Rheumatoid arthritis 549 HIV 911 Source: Murray and Lopez (1996)
Fuster, Voute. Lancet 2005
Road deaths, Europe (2004) 7000 6000 EC, Dept of Transport (The Times, 20.3.2006) 5000 4000 3000 TOTAL EU: 43,500 2000 1000 0 D I F E UK GR P NL
Outline Introduction What data sources are available? AMR-associated outcomes Current data about clinical and public health impact?
What are clinical implications of AMR? Treatment failure due to wrong choice Increased morbidity and mortality
Impact of antibiotic resistance on in-hospital mortality Pathogen OR 95 % CI P VRE 2.1 1.0-4.4.04 Pseudomonas spp 3.0 1.2-7.8.02 Enterobacter spp 5.0 1.1-22.9.01 Carmeli et al, Arch Intern Med 1999; 159: 1127-1132 Cosgrove et al, Arch Intern Med 2002; 162: 185-90 Carmeli et al, Arch Intern Med 2002; 162: 2223-2228
What are clinical implications of AMR? Treatment failure due to wrong choice Increased morbidity and mortality Use of more toxic, less efficacious and more expensive alternatives Example MRSA vs MSSA: Vancomycin/Linezolid vs. Oxacillin/Cephalosporins
What are clinical implications of AMR? Treatment failure due to wrong choice Increased morbidity and mortality Use of more toxic and less efficacious alternatives Vancomycin vs. Oxacillin Added burden of nosocomial infections
MSSA MRSA
Number of death certificates with MSSA/MRSA as underlying cause, UK 600 500 400 MRSA MSSA 300 200 100 0 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Health Statistics Quarterly, 2006
What are clinical implications of AMR? Treatment failure due to wrong choice Increased morbidity and mortality Use of more toxic alternatives Added burden of nosocomial infections Possibility of no alternate agents (e.g. VRSA, MDR-Tb, pan-resistant Stenotrophomonas or Acinetobacter spp)
Lundi 05/05/2003 (1)x1 (2)x1 (3)x1 ANTIBIOGRAMMES Smal MRSA MRSA Interpré. Interpré. Interpré. ----------------------------------------------------+ Penicilline G RESIST RESIST Flucloxacilline RESIST RESIST Piperacilline RESIST Piperac.+tazob RESIST ----------------------------------------------------+ Ceftazidime RESIST Cefepime RESIST Imipenem RESIST Aztreonam RESIST ----------------------------------------------------+ Amikacine RESIST RESIST Gentamicine RESIST RESIST RESIST Tobramycine RESIST ----------------------------------------------------+ Norfloxacine RESIST RESIST RESIST Ciprofloxacine RESIST RESIST RESIST ----------------------------------------------------+ Clindamycine RESIST RESIST Erythromycine RESIST RESIST ----------------------------------------------------+ Acide fusidique S S Co-trimoxazole RESIST S S Fosfomycine S S Rifampicine S S Vancomycine S S Teicoplanine S S
Mortality: pan-r S. maltophilia Crude mortality: 25% (10/40) Definite association with pan-r Stenotrophomonas maltophilia: n=1 (autopsy-proven pneumonia) Tsiodras et al, Scand J Infect Dis 2000; 32: 651-56
Outline Introduction What data sources are available? AMR-associated outcomes Current data about clinical and public health impact? Excess costs of MRSA?
Financial burden of MRSA Increased direct costs of providing care to MRSA-infected patients; Antibiotic treatment costs for therapy or empiric coverage of MRSA; Indirect costs & diminished quality of life; Infrastructure costs of surveillance programs and contact isolation.
MRSA Burden of isolation days Four University Hospitals 1 year (2002) MRSA adds at least per year: 21,665 isolation-days Median duration: 11 to 16 days 1.5% of all patient-days Chaberny et al, Int J Hyg Environ Health 2005; 208: 447-453
Daily cost: 372 Per case: 9260
Economic Impact of C-MRSA -- Children, 2001-2004, Texas -- Purcell et al. Ped Infect Dis J 2006; 25: 178-180 Per member per month expenses for cellulitis and abscess
Why is it so difficult to determine the impact and burden of disease of AMR?
Methodological Challenges -- Potential Confounders MICRO Pathogen Physician Patient / Host MACRO Population
Bacterial Pathogens: Virulence and AMR Resistance does not increase bacterial virulence (in general) Resistance may decrease: Bacterial survival fitness Transmissibility Virulence
CA-MRSA: Acquisition of resistance does not seem to decrease virulence!
Mortality and AMR in Gram-negative infections Harbarth et al; 1999 1,766 patients with GN-bacteremia Multi-resistance not associated with increased mortality Blot et al; 2002 Paterson et al; 2004 328 critically ill patients with GNbacteremia 440 patients with Klebsiella pneumoniae BSI No effect of AMR resistance on patient outcome Almost identical mortality (presence or absence of ESBL)
Impact of antimicrobial treatment appropriateness on patient outcome (904 cases of microbiologically documented severe sepsis) Harbarth et al. Am J Med 2003 Probability of survival (%) 100 75 50 25 0 Appropriate treatment Inappropriate treatment Adjusted OR = 1.8 (CI 95, 1.2-2.6) 0 10 20 30
Microbial etiology and treatment appropriateness in severe sepsis (904 cases of microbiologically documented severe sepsis -- LENERCEPT anti-tnf study) Harbarth et al. Am J Med 2003 - Adequately treated bacteria Streptococcus pneumoniae Escherichia coli Methicillin-sensitive S. aureus - Inadequately treated bacteria MRSA Pseudomonas aeruginosa Enterobacter spp Acinetobacter or Stenotrophomonas Inappropriate Antimicrobial Therapy (n = 211) 7 47 36 12 40 24 28 Appropriate Antimicrobial Therapy (n = 693) 130 176 105 8 42 32 11
PROBLEM High crude mortality in patients with infections caused by multidrug-resistant bacteria Drawbacks of crude mortality data: Failure to adjust for severity Failure to identify causally related death Most patients who die in the hospital die either with infection or because of infection
Excess mortality? of AMR-infections A No infection B AMR-infection Inevitable death? Accelerated death? Attributable death?
Excess morbidity (LOS and costs) Matched cohort study approach Prior Costs & LOS A CO B CA Infection Extra length of stay Extra costs
Time-varying exposures and matched cohort studies Source of bias: infected and uninfected patients are compared with regard to total hospital costs. Association between pre-infection costs and infection amplifies confounding. Matched cohort studies produce biased effect estimates and may overestimate excess costs. Samore & Harbarth. Chapter 93, Mayhall textbook. 3rd edition, 2004.
LOS related to S. aureus infection Type of infection MRSA LOS (d) MSSA LOS (d) Fold increase Attributable LOS (d) P SSI * 15 10 1.2 2.6.11 BSI 9 7 1.3 2.2.02 * Adjusted for comorbidity, hospital, ASA score, surgery, LOS before infection. Adjusted for dialysis, presence of prosthetic material, comorbidities, source, McCabe score. Engemann JJ et al. Clin Infect Dis 2003; 36: 592-98 Cosgrove S et al. Infect Control Hosp Epidemiol 2005
Charges comparing MRSA and multi-resistant GNB (Austria) (Median) MRSA (N = 74) MR-GNB (N = 99) Length of stay 37 42 Medical charges 6624 18,115 Unadjusted for comorbidity and severity of illness! Daxboeck et al. J Hosp Infect 2006; 62: 214-18
Conclusions Paucity of data regarding the impact of antimicrobial resistance on public health Consistency of data regarding the impact of antimicrobial resistance on clinical outcome MRSA & VRE Less evidence for PRP & ESBL Some evidence of: Increased likelihood of treatment failure Increased morbidity and mortality Added disease burden
Caveats Imputing increased mortality to resistance may be confounded by Greater severity of illness/comorbidity of patients who acquire resistant strains Inappropriate initial or subsequent therapy Research recommendation: Look at the timing of the events and rules of causality! Whatever the method chosen, make sure to adjust for underlying disease and adequacy of therapy!