- Email: [email protected]
Clinical impact of antimicrobial resistance: design matters – Authors' reply
Correspondence
Clinical impact of antimicrobial resistance: design matters Marie-Laurence Lambert and colleagues1 did a large retrospective study in intensive care units (ICU) from ten European countries. They concluded that ICU-acquired pneumonia and bloodstream infections substantially increased mortality and ICU stay, while the additional effect of antibiotic resistance in this group of patients was relatively modest. We believe that this study design is unlikely to fully capture the public health effect of antibiotic resistance for two reasons: first, patients with bacterial infection in ICUs often receive early empirical treatment that typically covers the resistance phenotypes that the authors chose as exposures; and second, the median follow-up was limited to 5 days, and this short follow-up might have concealed the long-term effects of treatment failure. The authors included a summary estimate for the effect of resistance in their discussion, thereby neglecting the pathogenspecific effect of antibiotic resistance on mortality. We have reported data for more than 2000 patients with Staphylococcus aureus or Escherichia coli bloodstream infections with follow-up beyond hospital discharge, and pathogen-specific estimates showed that resistance increased mortality 30-days after infection by 80–150%.2,3 In a study with 90-day Outcome
follow-up, the hazard of death in those with meticillin-resistant S aureus bacteraemia was twice that of meticillin-sensitive S aureus.4 We conclude that to address the burden of disease attributable to antibiotic resistance, comprehensive enrolment of patients is needed, including those in non-ICU settings, with follow-up beyond hospital discharge and a pathogen-specific approach to inform health-care providers and the public about the importance of this health threat. We declare that we have no conflicts of interest.
Hajo Grundmann, *Marlieke de Kraker, Peter Davey [email protected] National Institute for Public Health and the Environment, Centre for Infectious Disease Control, PO Box 1, Bilthoven, Netherlands (HG, MdK); Quality Safety and Informatics Research Group, The Mackenzie building, Kirsty Semple Way, Dundee DD2 4BF, Scotland, UK (PD) 1
2
3
4
Lambert ML, Suetens C, Savey A, et al. Clinical outcomes of health-care-associated infections and antimicrobial resistance in patients admitted to European intensive-care units: a cohort study. Lancet Infect Dis 2010; 11: 30–38. De Kraker MEA, Wolkewitz M, Davey PG, et al. The clinical impact of antimicrobial resistance in European hospitals: excess mortality and length of hospital stay related to methicillin resistant Staphylococcus aureus bloodstream infections. Antimicrob Agents Chemother 2011; 55: 1598–605. De Kraker MEA, Wolkewitz M, Davey PG, et al. Burden of antimicrobial resistance in European hospitals: excess mortality and length of hospital stay associated with third-generation cephalosporin resistant Escherichia coli bloodstream infections. J Antimicrob Chemother 2010; 66: 398–407. Wolkewitz M, Frank U, Philips G, et al. Mortality associated with in-hospital bacteraemia caused by Staphylococcus aureus: a multistate analysis with follow-up beyond hospital discharge. J Antimicrob Chemother 2010; 66: 381–86.
Authors’ reply We thank Hajo Grundmann and colleagues, our associated partners of the BURDEN project, for their comments; however, we would like to correct two factual inaccuracies. First, our study1 was not “neglecting the pathogen-specific effect of antibiotic resistance on mortality” because it provided two tables with pathogen-specific data for outcome. Second, the statement about short follow-up could be misunderstood. Although the median follow-up was 5 days (IQR 3–10 days), patients were followed up throughout their stay in the intensive-care unit (ICU). For patients who were discharged from the ICU alive, the median follow-up after bloodstream infection, for example, ranged from 9 days to 20 days depending on the microorganism and its resistance pattern.1 Another comment was that “patients with bacterial infection in ICUs often receive early empirical treatment that typically covers the resistance phenotypes that the authors chose as exposures”. Although this comment is true, we clearly stated that our results only applied to the most common patterns of resistance and that our study aimed to measure the real-life effect of these resistance patterns. Whether this effect was due to differences in appropriateness of treatment was beyond the scope of our study. We agree that the public health effects of antibiotic resistance and
Adjusted estimates for the burden of bloodstream infection and resistance* MRSA bloodstream infection vs no Staphylococcus aureus bloodstream infection OR or SHR (95% CI)
MSSA bloodstream infection vs no S aureus bloodstream infection OR or SHR (95% CI)
Burden of resistance: ratio of ORs and SHRs for MRSA vs MSSA
Time-matched subcohort of patients in hospitals in Europe2
Mortality 30 days after infection or enrolment
OR 4·4 (2·8–7·0)
OR 2·4 (1·7–3·3)
OR 1·8 (1·04–3·2)
Cohort of patients in intensive-care units in Europe1
Mortality in intensivecare unit
SHR 3·3 (2·5–5·2)
SHR 2·1 (1·6–2·6)
SHR 1·6 (1·1–2·3)
*Factors of adjustment differed between the two studies. MRSA=meticillin-resistant S aureus. MSSA=meticillin-susceptible S aureus. OR=odds ratio. SHR=subdistribution hazard ratio.
Table: Mortality related to Staphylococcus aureus bloodstream infection in two studies with different designs
344
www.thelancet.com/infection Vol 11 May 2011
Correspondence
disease should be studied at different levels of health care and not only in ICUs. The study by Marlieke de Kraker and colleagues2 adds important information about this effect at the hospital level. We compared the outcomes of Staphylococcus aureus bloodstream infections between our study and de Kraker and colleagues’ study.2 Both studies allowed for an exploration of the burden of bloodstream infections and for the additional burden of resistance and showed remarkably similar results (table). Therefore, the main difference between the two studies is not in their results, but in the interpretation of these results. Although de Kraker and colleagues emphasise resistance, we emphasised the high effect of the infections for two reasons: because a large proportion of these infections are preventable and because with no infection, there is no infection with resistant microorganisms. In our study, the effect of resistance was even smaller, and the effect of the infection was larger for other microorganisms—eg, Pseudomonas aeruginosa and Escherichia coli—than for S aureus. Infections with resistant microorganisms matter. So do infections with susceptible microorganisms. MW has received grant support and travel expenses from the Directorate General for Health and Consumer Protection. All other authors declare that they have no conflicts of interest.
*Marie-Laurence Lambert, Carl Suetens, Uwe Frank, Martin Wolkewitz [email protected] Healthcare-associated Infections Unit, Public Health and Surveillance Department, Scientific Institute for Public Health, Brussels 1050, Belgium (M-LL); Surveillance Unit, European Centre for Disease Prevention and Control, Stockholm, Sweden (CS); Department of Environmental Health Sciences, Freiburg University Medical Centre, Freiburg, Germany (UF); and Deptartment of Medical Biometry and Statistics, Institute of Medical Biometry and Medical Informatics, University Medical Centre Freiburg, Freiburg Germany (MW) 1
Lambert ML, Suetens C, Savey A, et al. Clinical outcomes of health-care-associated infections and antimicrobial resistance in patients admitted to European intensive-care units: a cohort study. Lancet Infect Dis 2011; 11: 30–38.
www.thelancet.com/infection Vol 11 May 2011
2
de Kraker MEA, Wolkewitz M, Davey PG, Grundmann H, for the BURDEN study group. The clinical impact of antimicrobial resistance in European hospitals: excess mortality and length of hospital stay related to methicillin resistant Staphylococcus aureus bloodstream infections. Antimicrob Agents Chemother 2011; published online Jan 10. DOI:10.1128/AAC.01157-10.
Schistosomiasis elimination Darren Gray and colleagues1 proposed a multifaceted integrated control programme for effective and sustainable control and elimination of schistosomiasis. We agree that a sustainable control strategy is necessary and urgent for many reasons, such as the rebound of the schistosomiasis epidemic, reduced financial support for schistosomiasisrelated research and work, reduced compliance rate for chemotherapy, snail dispersal, and frequent uncontrollable floods. However, where should the emphasis be for implementation of this control programme? The most effective way to eliminate such parasites is to interrupt transmission cycles. Schistosomiasis has only one intermediate host (Oncomelania hupensis). On the basis of results from successful projects in Japan and China, and simulation results from mathematical models, control of snail populations seems the most effective way to interrupt schistosomiasis transmission cycles for a sustainable control strategy.2–4 In regions where snail populations have been controlled or even eliminated (eg, Shanghai municipality, China), rebound of epidemics will not occur and reduced financial resources are not very important because the transmission cycle has been interrupted. However, snail habitats in the current epidemic regions are usually extensive, especially in lake and marshland regions. Large-scale control of these snail habitats, as
China did in the early 1950s, is no longer practical. Fortunately, the concept of active transmission sites provides us with a feasible method for effectively interrupting schistosomiasis transmission cycles. These sites are just a small portion of all snail habitats and are defined as “the high-risk snail habitats where infected snails are frequently present and with which people are often in contact”.5 Use of the strategy of snail control on active transmission sites, and other measures (eg, chemotherapy), in non-active transmission sites, will be practicable for gradual reduction in size of schistosomiasis high-risk regions and will help with sustainable control and interruption of transmission cycles. Although incorporation of the idea of active transmissions sites into the framework of integrated control programmes might be promising, more effective approaches to identify these sites deserve further research. Our research was supported by a grant from the National Important Technologies Project of China (grant number 2008ZX10004-011).
Zhijie Zhang, Qingwu Jiang [email protected], jiangqw@fudan. edu.cn Laboratory of Geographic Information and Spatial Analysis, Department of Geography, Faculty of Arts and Science, Queen’s University, 99 University Avenue, Kingston K7L 3N6, Ontario, Canada (ZZ); Department of Epidemiology, School of Public Health, Fudan University,No 138 Yi Xue Yuan Road, Shanghai 200032, China (ZZ, QJ) 1
2
3
4
5
Gray DJ, McManus DP, Li Y, Williams GM, Bergquist R, Ross AG. Schistosomiasis elimination: lessons from the past guide the future. Lancet Infect Dis 2010; 10: 733–36. Chen XY, Wang LY, Cai JM, et al. Schistosomiasis control in China: the impact of a 10-year World Bank Loan Project (1992–2001). Bull World Health Organ 2005; 83: 43–48. Tanaka H, Tsuji M. From discovery to eradication of schistosomiasis in Japan: 1847–1996. Int J Parasitol 1997; 27: 1465–80. Zhang ZJ, Ong SH, Lynn HS, et al. Generalized negative binomial distribution: a promising statistical distribution for Oncomelania hupensis in the lake- and marsh-land regions of China. Ann Trop Med Parasitol 2008; 102: 541–52. Zhang ZJ, Carpenter TE, Lynn HS, et al. Location of active transmission sites of Schistosoma japonicum in lake and marshland regions in China. Parasitology 2009; 136: 737–46.
345
Clinical impact of antimicrobial resistance: design matters Marie-Laurence Lambert and colleagues1 did a large retrospective study in intensive care units (ICU) from ten European countries. They concluded that ICU-acquired pneumonia and bloodstream infections substantially increased mortality and ICU stay, while the additional effect of antibiotic resistance in this group of patients was relatively modest. We believe that this study design is unlikely to fully capture the public health effect of antibiotic resistance for two reasons: first, patients with bacterial infection in ICUs often receive early empirical treatment that typically covers the resistance phenotypes that the authors chose as exposures; and second, the median follow-up was limited to 5 days, and this short follow-up might have concealed the long-term effects of treatment failure. The authors included a summary estimate for the effect of resistance in their discussion, thereby neglecting the pathogenspecific effect of antibiotic resistance on mortality. We have reported data for more than 2000 patients with Staphylococcus aureus or Escherichia coli bloodstream infections with follow-up beyond hospital discharge, and pathogen-specific estimates showed that resistance increased mortality 30-days after infection by 80–150%.2,3 In a study with 90-day Outcome
follow-up, the hazard of death in those with meticillin-resistant S aureus bacteraemia was twice that of meticillin-sensitive S aureus.4 We conclude that to address the burden of disease attributable to antibiotic resistance, comprehensive enrolment of patients is needed, including those in non-ICU settings, with follow-up beyond hospital discharge and a pathogen-specific approach to inform health-care providers and the public about the importance of this health threat. We declare that we have no conflicts of interest.
Hajo Grundmann, *Marlieke de Kraker, Peter Davey [email protected] National Institute for Public Health and the Environment, Centre for Infectious Disease Control, PO Box 1, Bilthoven, Netherlands (HG, MdK); Quality Safety and Informatics Research Group, The Mackenzie building, Kirsty Semple Way, Dundee DD2 4BF, Scotland, UK (PD) 1
2
3
4
Lambert ML, Suetens C, Savey A, et al. Clinical outcomes of health-care-associated infections and antimicrobial resistance in patients admitted to European intensive-care units: a cohort study. Lancet Infect Dis 2010; 11: 30–38. De Kraker MEA, Wolkewitz M, Davey PG, et al. The clinical impact of antimicrobial resistance in European hospitals: excess mortality and length of hospital stay related to methicillin resistant Staphylococcus aureus bloodstream infections. Antimicrob Agents Chemother 2011; 55: 1598–605. De Kraker MEA, Wolkewitz M, Davey PG, et al. Burden of antimicrobial resistance in European hospitals: excess mortality and length of hospital stay associated with third-generation cephalosporin resistant Escherichia coli bloodstream infections. J Antimicrob Chemother 2010; 66: 398–407. Wolkewitz M, Frank U, Philips G, et al. Mortality associated with in-hospital bacteraemia caused by Staphylococcus aureus: a multistate analysis with follow-up beyond hospital discharge. J Antimicrob Chemother 2010; 66: 381–86.
Authors’ reply We thank Hajo Grundmann and colleagues, our associated partners of the BURDEN project, for their comments; however, we would like to correct two factual inaccuracies. First, our study1 was not “neglecting the pathogen-specific effect of antibiotic resistance on mortality” because it provided two tables with pathogen-specific data for outcome. Second, the statement about short follow-up could be misunderstood. Although the median follow-up was 5 days (IQR 3–10 days), patients were followed up throughout their stay in the intensive-care unit (ICU). For patients who were discharged from the ICU alive, the median follow-up after bloodstream infection, for example, ranged from 9 days to 20 days depending on the microorganism and its resistance pattern.1 Another comment was that “patients with bacterial infection in ICUs often receive early empirical treatment that typically covers the resistance phenotypes that the authors chose as exposures”. Although this comment is true, we clearly stated that our results only applied to the most common patterns of resistance and that our study aimed to measure the real-life effect of these resistance patterns. Whether this effect was due to differences in appropriateness of treatment was beyond the scope of our study. We agree that the public health effects of antibiotic resistance and
Adjusted estimates for the burden of bloodstream infection and resistance* MRSA bloodstream infection vs no Staphylococcus aureus bloodstream infection OR or SHR (95% CI)
MSSA bloodstream infection vs no S aureus bloodstream infection OR or SHR (95% CI)
Burden of resistance: ratio of ORs and SHRs for MRSA vs MSSA
Time-matched subcohort of patients in hospitals in Europe2
Mortality 30 days after infection or enrolment
OR 4·4 (2·8–7·0)
OR 2·4 (1·7–3·3)
OR 1·8 (1·04–3·2)
Cohort of patients in intensive-care units in Europe1
Mortality in intensivecare unit
SHR 3·3 (2·5–5·2)
SHR 2·1 (1·6–2·6)
SHR 1·6 (1·1–2·3)
*Factors of adjustment differed between the two studies. MRSA=meticillin-resistant S aureus. MSSA=meticillin-susceptible S aureus. OR=odds ratio. SHR=subdistribution hazard ratio.
Table: Mortality related to Staphylococcus aureus bloodstream infection in two studies with different designs
344
www.thelancet.com/infection Vol 11 May 2011
Correspondence
disease should be studied at different levels of health care and not only in ICUs. The study by Marlieke de Kraker and colleagues2 adds important information about this effect at the hospital level. We compared the outcomes of Staphylococcus aureus bloodstream infections between our study and de Kraker and colleagues’ study.2 Both studies allowed for an exploration of the burden of bloodstream infections and for the additional burden of resistance and showed remarkably similar results (table). Therefore, the main difference between the two studies is not in their results, but in the interpretation of these results. Although de Kraker and colleagues emphasise resistance, we emphasised the high effect of the infections for two reasons: because a large proportion of these infections are preventable and because with no infection, there is no infection with resistant microorganisms. In our study, the effect of resistance was even smaller, and the effect of the infection was larger for other microorganisms—eg, Pseudomonas aeruginosa and Escherichia coli—than for S aureus. Infections with resistant microorganisms matter. So do infections with susceptible microorganisms. MW has received grant support and travel expenses from the Directorate General for Health and Consumer Protection. All other authors declare that they have no conflicts of interest.
*Marie-Laurence Lambert, Carl Suetens, Uwe Frank, Martin Wolkewitz [email protected] Healthcare-associated Infections Unit, Public Health and Surveillance Department, Scientific Institute for Public Health, Brussels 1050, Belgium (M-LL); Surveillance Unit, European Centre for Disease Prevention and Control, Stockholm, Sweden (CS); Department of Environmental Health Sciences, Freiburg University Medical Centre, Freiburg, Germany (UF); and Deptartment of Medical Biometry and Statistics, Institute of Medical Biometry and Medical Informatics, University Medical Centre Freiburg, Freiburg Germany (MW) 1
Lambert ML, Suetens C, Savey A, et al. Clinical outcomes of health-care-associated infections and antimicrobial resistance in patients admitted to European intensive-care units: a cohort study. Lancet Infect Dis 2011; 11: 30–38.
www.thelancet.com/infection Vol 11 May 2011
2
de Kraker MEA, Wolkewitz M, Davey PG, Grundmann H, for the BURDEN study group. The clinical impact of antimicrobial resistance in European hospitals: excess mortality and length of hospital stay related to methicillin resistant Staphylococcus aureus bloodstream infections. Antimicrob Agents Chemother 2011; published online Jan 10. DOI:10.1128/AAC.01157-10.
Schistosomiasis elimination Darren Gray and colleagues1 proposed a multifaceted integrated control programme for effective and sustainable control and elimination of schistosomiasis. We agree that a sustainable control strategy is necessary and urgent for many reasons, such as the rebound of the schistosomiasis epidemic, reduced financial support for schistosomiasisrelated research and work, reduced compliance rate for chemotherapy, snail dispersal, and frequent uncontrollable floods. However, where should the emphasis be for implementation of this control programme? The most effective way to eliminate such parasites is to interrupt transmission cycles. Schistosomiasis has only one intermediate host (Oncomelania hupensis). On the basis of results from successful projects in Japan and China, and simulation results from mathematical models, control of snail populations seems the most effective way to interrupt schistosomiasis transmission cycles for a sustainable control strategy.2–4 In regions where snail populations have been controlled or even eliminated (eg, Shanghai municipality, China), rebound of epidemics will not occur and reduced financial resources are not very important because the transmission cycle has been interrupted. However, snail habitats in the current epidemic regions are usually extensive, especially in lake and marshland regions. Large-scale control of these snail habitats, as
China did in the early 1950s, is no longer practical. Fortunately, the concept of active transmission sites provides us with a feasible method for effectively interrupting schistosomiasis transmission cycles. These sites are just a small portion of all snail habitats and are defined as “the high-risk snail habitats where infected snails are frequently present and with which people are often in contact”.5 Use of the strategy of snail control on active transmission sites, and other measures (eg, chemotherapy), in non-active transmission sites, will be practicable for gradual reduction in size of schistosomiasis high-risk regions and will help with sustainable control and interruption of transmission cycles. Although incorporation of the idea of active transmissions sites into the framework of integrated control programmes might be promising, more effective approaches to identify these sites deserve further research. Our research was supported by a grant from the National Important Technologies Project of China (grant number 2008ZX10004-011).
Zhijie Zhang, Qingwu Jiang [email protected], jiangqw@fudan. edu.cn Laboratory of Geographic Information and Spatial Analysis, Department of Geography, Faculty of Arts and Science, Queen’s University, 99 University Avenue, Kingston K7L 3N6, Ontario, Canada (ZZ); Department of Epidemiology, School of Public Health, Fudan University,No 138 Yi Xue Yuan Road, Shanghai 200032, China (ZZ, QJ) 1
2
3
4
5
Gray DJ, McManus DP, Li Y, Williams GM, Bergquist R, Ross AG. Schistosomiasis elimination: lessons from the past guide the future. Lancet Infect Dis 2010; 10: 733–36. Chen XY, Wang LY, Cai JM, et al. Schistosomiasis control in China: the impact of a 10-year World Bank Loan Project (1992–2001). Bull World Health Organ 2005; 83: 43–48. Tanaka H, Tsuji M. From discovery to eradication of schistosomiasis in Japan: 1847–1996. Int J Parasitol 1997; 27: 1465–80. Zhang ZJ, Ong SH, Lynn HS, et al. Generalized negative binomial distribution: a promising statistical distribution for Oncomelania hupensis in the lake- and marsh-land regions of China. Ann Trop Med Parasitol 2008; 102: 541–52. Zhang ZJ, Carpenter TE, Lynn HS, et al. Location of active transmission sites of Schistosoma japonicum in lake and marshland regions in China. Parasitology 2009; 136: 737–46.
345