Eur J Clin Microbiol Infect Dis DOI 10.1007/s10096-014-2114-y
ARTICLE
Prior infections are associated with increased mortality from subsequent blood-stream infections among patients with hematological malignancies J. A. Greenberg & M. Z. David & D. L. Pitrak & J. B. Hall & J. P. Kress
Received: 18 January 2014 / Accepted: 7 April 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Many patients who are evaluated and treated for sepsis have histories of recent infections. The prognostic implications of surviving an infectious process are not well understood. We undertook this study to determine the clinical impact of prior infections among patients with hematological malignancies, a population at high risk for developing and dying from sepsis. The medical records of 203 patients with hematological malignancies and bloodstream infections admitted over a 3-year period to an urban teaching hospital were retrospectively reviewed. The 30-day mortality after blood-stream infection in these high-risk patients was 24 %. There were 46 patients (23 %) who had inpatient infections in the 90 days prior to the index blood-stream infection. History of recent infection portended worse prognosis from blood-stream infection under multivariable analysis [odds ratio (OR) 2.60, p=0.04, 95 % confidence interval (CI) 1.04–6.47]. There were 86 patients (42 %) who had subsequent infections in the first 90 days after the index blood-stream infection. Patients with subsequent infections had greater mortality during days 91–365 than patients without subsequent infections [hazard ratio (HR) 1.97, p=0.02, 95 % CI 1.13–3.44]. Recent infections prognosticate worse outcomes from subsequent blood-stream infections for this high-risk population. Further research into the clinical and biochemical reasons for this observation may lead to
J. A. Greenberg (*) : J. B. Hall : J. P. Kress Section of Pulmonary and Critical Care Medicine, Department of Medicine, University of Chicago, Chicago, IL, USA e-mail: [email protected] M. Z. David : D. L. Pitrak Section of Infectious Disease and Global Health, Department of Medicine, University of Chicago, Chicago, IL, USA
targets for intervention, and, ultimately, improvements in long-term mortality from sepsis.
Introduction Patients with hematological malignancies are among the highest risk of developing and dying from severe infections [1]. Of the broad range of infectious processes for which these severely immunocompromised patients are at risk, blood-stream infections may be the most common [2–4]. In a population-based study in Denmark, 21 % of patients with newly diagnosed hematological malignancies developed bacteremia during a median follow up of 1.1 years [5]. In a hospital surveillance study, the incidence of blood-stream infections ranged from 12.5 per 1,000 patient-days for chronic myelogenous leukemia to 30.5 per 1,000 patient-days for acute myelogenous leukemia [4]. Often, these infections are costly, delay chemotherapy, and have considerable mortality, with 16–32 % of patients dying within 30 days of a blood-stream infection [6–8]. It is unclear the degree to which characteristics of the pathogen, the host’s underlying malignancy, and the host’s clinical response to infection affect patient outcome. Therefore, we aimed to thoroughly characterize hematological malignancy and infection-related variables to determine the most influential predictors of 30-day mortality after a blood-stream infection. In addition, given the fact that members of this population are at high risk for multiple infectious processes, we hypothesized that surviving a recent infection could be a prognostic factor for mortality from a subsequent blood-stream infection.
Eur J Clin Microbiol Infect Dis
Materials and methods Study location and patient population This study was conducted at the University of Chicago Medical Center in Chicago, Illinois, a 547-bed, universityaffiliated, urban teaching hospital. The University of Chicago Institutional Review Board approved this study and waived the need to obtain informed consent. A microbiology database of all positive blood cultures from inpatients from April 1, 2009 to April 1, 2012 was obtained. The hospital charts for patients who met the criteria for blood-stream infections based on the United States Centers for Disease Control and Prevention (CDC) definition [9] were manually reviewed. Patients were included if they had a diagnosis of a hematological malignancy and were ≥18 years of age. Patients were excluded if: (1) their index blood-stream infection occurred within 90 days of a previous blood-stream infection occurring before April 1, 2009; (2) their index bloodstream infection occurred within 90 days of a previous bloodstream infection occurring at an outside institution; (3) it was unknown as to whether they were alive or deceased at 30 days after blood-stream infection. Subsequent blood-stream infections for a given patient were not considered an index bloodstream infection. Study design and data collection This study used a retrospective cohort design. All data were collected by review of the patients’ electronic medical records, which included nursing documentation, laboratory values, microbiology data, physician notes, radiology reports, and hospital discharge summaries. The primary end point for patients with blood-stream infections was 30-day mortality. The secondary end point was mortality 91–365 days after blood-stream infection. When mortality could not be determined from the medical records, the Social Security Death Index was reviewed. The recorded baseline characteristics were age, gender, race, Charlson comorbidity index [10], type of hematological malignancy, and variables related to malignancy treatment course. A high-grade malignancy was defined as acute leukemia or high-grade lymphoma. Patients who were unable to independently perform activities of daily living on hospital admission were defined as having poor performance status. Neutropenia was defined as an absolute neutrophil count less than 500 cells per microliter. All inpatient infections in the 90 days prior to blood-stream infection were recorded. Designation of an infectious process was based on the United States CDC definition [9] and an international consensus definition for fungal infections [11]. Infections treated at outside institutions were also included if documented in their medical records.
The following characteristics of the blood-stream infections were recorded: hospital location of the patient when first positive blood cultures were drawn, source of infection, and whether cultures were positive for more than one day (persistently positive cultures). A bacterial organism was labeled as antibioticresistant if it was methicillin-resistant Staphylococcus aureus, vancomycin-resistant Enterococcus, or a Gram-negative bacteria resistant to cefepime, piperacillin–tazobactam, or a carbapenem based on microbiology laboratory susceptibility testing. The initial antibiotic choice was deemed appropriate if the organism was susceptible under in vitro testing and the antibiotic was given within 24 h of positive blood cultures. The number of positive systemic inflammatory response criteria was recorded [12]. The severity of illness was determined by the Sequential Organ Failure Assessment (SOFA) score, which was calculated on the first day of positive blood cultures [13]. The first subsequent infection over days 1–90 was recorded using the previously referenced consensus definitions [9, 11]. Data analysis Continuous variables were reported as medians with 25th and 75th percentiles. The Student’s t-test was used when comparing distributions of normally distributed data, and the Mann– Whitney U test was employed to analyze non-normally distributed data. Categorical data were expressed as frequency distributions, and either the Chi-squared test or the Fisher’s exact test were used to determine if differences existed between groups. Logistic regression analysis was used to identify risk factors associated with 30-day mortality. An initial model included all recorded variables; a backward selection procedure was then employed to determine a final model. Kaplan–Meier survival functions and a Cox proportional hazards model were used to investigate factors associated with mortality up to 365 days after blood-stream infection. All tests were two-sided and a p-value ≤0.05 was considered to indicate statistical significance. All analyses were performed with Stata 12.1 (StataCorp, College Station, TX).
Results Between April 1, 2009 and April 1, 2012, patients with hematological malignancies had 306 blood-stream infections. There were three patients who were excluded because they had blood-stream infections at outside institutions in the 90 days prior to the index blood-stream infection. There were four patients who were excluded because they had bloodstream infections before the enrollment date of April 1, 2009 that were within 90 days of the index blood-stream infection. There were five patients with unknown dispositions at 30 days who were excluded. Ninety-one blood-stream infections were not counted as index blood-stream infections because they
Eur J Clin Microbiol Infect Dis
were subsequent infections in patients already included for analysis. Thus, 203 patients were included in the analysis. There were 49 patients (24 %) who met the primary end point of 30-day mortality. There were 69 patients (34 %) who required intensive care unit admission. Table 1 displays the bivariate associations with 30-day mortality for all recorded variables. Among the demographic variables, greater age was associated with higher mortality. Among malignancy-related variables, poor performance status on hospital admission and greater time since malignancy diagnosis were associated with increased mortality. A high-grade malignancy, neutropenia on the day of first positive blood cultures, and treatment with chemotherapy in the previous 90 days were associated with improved mortality. Among the characteristics of the index blood-stream infection, persistently positive cultures and more severe acute illness as measured by the SOFA score were associated with increased mortality. Blood-stream infection with Streptococcus species was associated with improved mortality compared to non-Streptococcal blood-stream infections. Finally, there were 46 patients (23 %) with inpatient infections in the 90 days prior to the index blood-stream infection; in the bivariate analysis, patients in this group had an increased risk of death after the index blood-stream infection [odds ratio (OR) 2.98, p
ARTICLE
Prior infections are associated with increased mortality from subsequent blood-stream infections among patients with hematological malignancies J. A. Greenberg & M. Z. David & D. L. Pitrak & J. B. Hall & J. P. Kress
Received: 18 January 2014 / Accepted: 7 April 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Many patients who are evaluated and treated for sepsis have histories of recent infections. The prognostic implications of surviving an infectious process are not well understood. We undertook this study to determine the clinical impact of prior infections among patients with hematological malignancies, a population at high risk for developing and dying from sepsis. The medical records of 203 patients with hematological malignancies and bloodstream infections admitted over a 3-year period to an urban teaching hospital were retrospectively reviewed. The 30-day mortality after blood-stream infection in these high-risk patients was 24 %. There were 46 patients (23 %) who had inpatient infections in the 90 days prior to the index blood-stream infection. History of recent infection portended worse prognosis from blood-stream infection under multivariable analysis [odds ratio (OR) 2.60, p=0.04, 95 % confidence interval (CI) 1.04–6.47]. There were 86 patients (42 %) who had subsequent infections in the first 90 days after the index blood-stream infection. Patients with subsequent infections had greater mortality during days 91–365 than patients without subsequent infections [hazard ratio (HR) 1.97, p=0.02, 95 % CI 1.13–3.44]. Recent infections prognosticate worse outcomes from subsequent blood-stream infections for this high-risk population. Further research into the clinical and biochemical reasons for this observation may lead to
J. A. Greenberg (*) : J. B. Hall : J. P. Kress Section of Pulmonary and Critical Care Medicine, Department of Medicine, University of Chicago, Chicago, IL, USA e-mail: [email protected] M. Z. David : D. L. Pitrak Section of Infectious Disease and Global Health, Department of Medicine, University of Chicago, Chicago, IL, USA
targets for intervention, and, ultimately, improvements in long-term mortality from sepsis.
Introduction Patients with hematological malignancies are among the highest risk of developing and dying from severe infections [1]. Of the broad range of infectious processes for which these severely immunocompromised patients are at risk, blood-stream infections may be the most common [2–4]. In a population-based study in Denmark, 21 % of patients with newly diagnosed hematological malignancies developed bacteremia during a median follow up of 1.1 years [5]. In a hospital surveillance study, the incidence of blood-stream infections ranged from 12.5 per 1,000 patient-days for chronic myelogenous leukemia to 30.5 per 1,000 patient-days for acute myelogenous leukemia [4]. Often, these infections are costly, delay chemotherapy, and have considerable mortality, with 16–32 % of patients dying within 30 days of a blood-stream infection [6–8]. It is unclear the degree to which characteristics of the pathogen, the host’s underlying malignancy, and the host’s clinical response to infection affect patient outcome. Therefore, we aimed to thoroughly characterize hematological malignancy and infection-related variables to determine the most influential predictors of 30-day mortality after a blood-stream infection. In addition, given the fact that members of this population are at high risk for multiple infectious processes, we hypothesized that surviving a recent infection could be a prognostic factor for mortality from a subsequent blood-stream infection.
Eur J Clin Microbiol Infect Dis
Materials and methods Study location and patient population This study was conducted at the University of Chicago Medical Center in Chicago, Illinois, a 547-bed, universityaffiliated, urban teaching hospital. The University of Chicago Institutional Review Board approved this study and waived the need to obtain informed consent. A microbiology database of all positive blood cultures from inpatients from April 1, 2009 to April 1, 2012 was obtained. The hospital charts for patients who met the criteria for blood-stream infections based on the United States Centers for Disease Control and Prevention (CDC) definition [9] were manually reviewed. Patients were included if they had a diagnosis of a hematological malignancy and were ≥18 years of age. Patients were excluded if: (1) their index blood-stream infection occurred within 90 days of a previous blood-stream infection occurring before April 1, 2009; (2) their index bloodstream infection occurred within 90 days of a previous bloodstream infection occurring at an outside institution; (3) it was unknown as to whether they were alive or deceased at 30 days after blood-stream infection. Subsequent blood-stream infections for a given patient were not considered an index bloodstream infection. Study design and data collection This study used a retrospective cohort design. All data were collected by review of the patients’ electronic medical records, which included nursing documentation, laboratory values, microbiology data, physician notes, radiology reports, and hospital discharge summaries. The primary end point for patients with blood-stream infections was 30-day mortality. The secondary end point was mortality 91–365 days after blood-stream infection. When mortality could not be determined from the medical records, the Social Security Death Index was reviewed. The recorded baseline characteristics were age, gender, race, Charlson comorbidity index [10], type of hematological malignancy, and variables related to malignancy treatment course. A high-grade malignancy was defined as acute leukemia or high-grade lymphoma. Patients who were unable to independently perform activities of daily living on hospital admission were defined as having poor performance status. Neutropenia was defined as an absolute neutrophil count less than 500 cells per microliter. All inpatient infections in the 90 days prior to blood-stream infection were recorded. Designation of an infectious process was based on the United States CDC definition [9] and an international consensus definition for fungal infections [11]. Infections treated at outside institutions were also included if documented in their medical records.
The following characteristics of the blood-stream infections were recorded: hospital location of the patient when first positive blood cultures were drawn, source of infection, and whether cultures were positive for more than one day (persistently positive cultures). A bacterial organism was labeled as antibioticresistant if it was methicillin-resistant Staphylococcus aureus, vancomycin-resistant Enterococcus, or a Gram-negative bacteria resistant to cefepime, piperacillin–tazobactam, or a carbapenem based on microbiology laboratory susceptibility testing. The initial antibiotic choice was deemed appropriate if the organism was susceptible under in vitro testing and the antibiotic was given within 24 h of positive blood cultures. The number of positive systemic inflammatory response criteria was recorded [12]. The severity of illness was determined by the Sequential Organ Failure Assessment (SOFA) score, which was calculated on the first day of positive blood cultures [13]. The first subsequent infection over days 1–90 was recorded using the previously referenced consensus definitions [9, 11]. Data analysis Continuous variables were reported as medians with 25th and 75th percentiles. The Student’s t-test was used when comparing distributions of normally distributed data, and the Mann– Whitney U test was employed to analyze non-normally distributed data. Categorical data were expressed as frequency distributions, and either the Chi-squared test or the Fisher’s exact test were used to determine if differences existed between groups. Logistic regression analysis was used to identify risk factors associated with 30-day mortality. An initial model included all recorded variables; a backward selection procedure was then employed to determine a final model. Kaplan–Meier survival functions and a Cox proportional hazards model were used to investigate factors associated with mortality up to 365 days after blood-stream infection. All tests were two-sided and a p-value ≤0.05 was considered to indicate statistical significance. All analyses were performed with Stata 12.1 (StataCorp, College Station, TX).
Results Between April 1, 2009 and April 1, 2012, patients with hematological malignancies had 306 blood-stream infections. There were three patients who were excluded because they had blood-stream infections at outside institutions in the 90 days prior to the index blood-stream infection. There were four patients who were excluded because they had bloodstream infections before the enrollment date of April 1, 2009 that were within 90 days of the index blood-stream infection. There were five patients with unknown dispositions at 30 days who were excluded. Ninety-one blood-stream infections were not counted as index blood-stream infections because they
Eur J Clin Microbiol Infect Dis
were subsequent infections in patients already included for analysis. Thus, 203 patients were included in the analysis. There were 49 patients (24 %) who met the primary end point of 30-day mortality. There were 69 patients (34 %) who required intensive care unit admission. Table 1 displays the bivariate associations with 30-day mortality for all recorded variables. Among the demographic variables, greater age was associated with higher mortality. Among malignancy-related variables, poor performance status on hospital admission and greater time since malignancy diagnosis were associated with increased mortality. A high-grade malignancy, neutropenia on the day of first positive blood cultures, and treatment with chemotherapy in the previous 90 days were associated with improved mortality. Among the characteristics of the index blood-stream infection, persistently positive cultures and more severe acute illness as measured by the SOFA score were associated with increased mortality. Blood-stream infection with Streptococcus species was associated with improved mortality compared to non-Streptococcal blood-stream infections. Finally, there were 46 patients (23 %) with inpatient infections in the 90 days prior to the index blood-stream infection; in the bivariate analysis, patients in this group had an increased risk of death after the index blood-stream infection [odds ratio (OR) 2.98, p
