Abstract
Background
There are concerns that emerging resistance to fluoroquinolones (FQ) may be leading to increasing rates of gram-negative rod (GNR) bacteremia in hematopoietic cell transplant (HCT) recipients. We set out to describe time trends in the incidence rates (IR) of GNR bacteremia and FQ-resistant GNR bacteremia in HCT recipients during an era of levofloxacin prophylaxis.
Methods
We conducted a longitudinal retrospective study of adults undergoing allogeneic HCT between 2003 and 2012 at the Seattle Cancer Care Alliance (SCCA). Annual trends in the IRs of GNR bacteremia and FQ-resistant GNR bacteremia through 100 days post-transplant were assessed using Poisson regression. Cox proportional hazards regression was used to compare 30-day mortality between patients with FQ-resistant and those with FQ-sensitive GNR bacteremia.
Results
Of the 2306 patients included in this cohort, 280 (12.1%) had GNR bacteremia. The IRs of GNR bacteremia and FQ-resistant GNR bacteremia increased from 2003 to 2009 and decreased afterwards; however, the overall annual trends were not significant (Incidence rate ratio [IRR] =1.01; 95% confidence interval [CI]: 0.98, 1.05 and IRR=1.01; 95% CI: 0.95, 1.08, respectively). FQ-resistant GNR bacteremia was associated with increased mortality compared to FQ-sensitive GNR bacteremia, even after adjustment for underlying disease severity, conditioning regimen, and age at transplant (Hazard ratio=2.11, 95% CI: 1.06, 4.23).
Conclusions
On average, rates of FQ-resistant GNR bacteremia have not significantly changed at the SCCA over 10 years of FQ prophylaxis, although FQ-resistant GNR bacteremia is associated with increased mortality compared to FQ-sensitive GNR bacteremia.
Keywords: Hematopoietic cell transplant, fluoroquinolone resistance, gram-negative
INTRODUCTION
Bacterial bloodborne infections are a common cause of morbidity and mortality among allogeneic hematopoietic cell transplant (HCT) patients, occurring in up to 55% of this population.1 Gram-negative rod (GNR) bacteremia affects between 5% and 11% of patients, and is associated with higher case-fatality than bacteremia caused by gram-positive organisms.2–4 The majority of cancer/transplant centers worldwide utilize antibiotic prophylaxis during neutropenia to prevent GNR bacteremia and associated mortality in these highly immunocompromised patients.
International guidelines currently recommend the use of fluoroquinolones (FQ) for neutropenia prophylaxis in cancer and HCT patients whose absolute neutrophil count is anticipated to decrease to ≤ 500/μL for at least seven days,5–7 as they have been shown to reduce the incidence of infectious complications in neutropenic patients and to decrease mortality.8 Levofloxacin is the most frequently used agent due to its excellent bioavailability, oral formulation, and the convenience of once daily dosing.9 While FQ prophylaxis has given flexibility to outpatient cancer care, recent guidelines warn that resistance should be monitored due to increasing FQ resistance worldwide.5 Recent studies have suggested that the incidence of GNR bacteremia is increasing in HCT recipients,1,10 including one study at our center.11 Some have speculated that increasing FQ use and development of associated antimicrobial resistance may play a major role in this change.10 However, studies that have examined the emergence of FQ-resistant GNR bacteremia among populations who receive FQ prophylaxis have been inconsistent.1,10,12 It is also unknown if HCT recipients who develop FQ-resistant GNR bacteremia have an associated increased mortality when compared to patients who develop FQ-sensitive GNR bacteremia. Data from other populations have demonstrated higher mortality associated with antimicrobial resistant infections,13 but most studies in HCT recipients have had insufficient power to address this important question.
Levofloxacin became standard practice for neutropenic prophylaxis for adult HCT recipients at the Seattle Cancer Care Alliance (SCCA)/Fred Hutchinson Cancer Research Center (FHCRC) in August 2002. The short-term impact of the change from ceftazidime to levofloxacin prophylaxis was addressed in a previous study at our center,9 but data on GNR infections after long-term use of levofloxacin prophylaxis as the standard of care have yet to be examined. In order to better understand trends in GNR bacteremia and FQ-resistant GNR bacteremia during this era of levofloxacin prophylaxis, we conducted a longitudinal retrospective study of allogeneic HCT recipients who were transplanted between January 2003 and December 2012. Our primary goal was to determine annual trends in the incidence of GNR bacteremia and FQ-resistant GNR bacteremia during the first 100 days post-transplant in this cohort. In addition, we compared 30-day mortality between HCT recipients with FQ-resistant GNR bacteremia and those with FQ-sensitive GNR bacteremia.
METHODS
Study Population
All adults (≥18 years) who underwent an allogeneic HCT between 1/1/2003 and 12/31/2012 at the SCCA/FHCRC were eligible for inclusion in this study. For those patients who had multiple allogeneic transplants during the study period of interest, each transplant was considered separately. The study was approved by the FHCRC Institutional Review Board.
Microbiologic assessment and antibacterial prophylaxis
Center guidelines recommend that two sets of anaerobic and aerobic blood cultures are drawn when a patient presents with a fever, followed by daily blood cultures until an alternative source of the fever is identified or the patient defervesces. Although standard recommendations exist, blood cultures are ultimately drawn at the discretion of the healthcare teams, with the exception of surveillance blood cultures that are routinely drawn per protocol from patients who are treated with high dose glucocorticoids (≥0.5mg/kg). Since steroids are known to blunt febrile responses,14 these surveillance cultures are drawn bi-weekly while inpatient, weekly on outpatient discharge, and are discontinued following tapering of glucocorticoids to <0.5mg/kg. The majority of blood cultures are drawn through central venous catheters or ports during post-HCT care.
HCT recipients at the SCCA/FHCRC receive 750 mg levofloxacin daily for prophylaxis at the start of neutropenia until neutrophil recovery (ANC >500/μL); levofloxacin is re-started for patients whose ANC drops below 500/μL at other points during their post-transplant care. Ceftazidime is the recommended empiric first-line antibiotic for neutropenic fever; the addition of vancomycin is only recommended if febrile patients also have high-grade mucositis. All patients also receive antimicrobial prophylaxis with trimethoprim-sulfamethoxazole (TMP), dapsone, or atovaquone for Pneumocystis jirovecii prophylaxis following ANC recovery, fluconazole or an extended spectrum azole (voriconazole or posconazole) for antifungal prophylaxis, and acyclovir or valacyclovir for herpes simplex/varicella zoster virus prophylaxis. All patients undergo cytomegalovirus (CMV) preemptive surveillance/therapy as has been previously described.11 Of note, the standard preemptive CMV surveillance strategy changed from pp65 antigenemia to DNA measurement by quantitative PCR in 2007.
Data collection
Data were extracted from prospectively collected databases maintained by the FHCRC that include demographic, laboratory, and clinical data from all patients undergoing HCT. Additional microbiologic data were collected through electronic medical record review. Allogeneic HCT recipients remain at the center for a minimum of 100 days post-transplant, assuring complete post-transplant data capture during this time period.
Definitions
GNR bacteremia was defined as the isolation of any GNR from a blood culture. To reduce the likelihood of misclassifying repeat blood cultures from a primary event, positive cultures for the same organism collected ≤ 14 days from a prior positive culture were considered part of the primary event. Additionally, early post-transplant GNR events were excluded if the same organism was isolated in a pre-transplant culture in a similar 14-day window. Positive cultures for different bacterial species, even if they occurred within 14 days from a documented GNR event, were considered unique events. Cultures that isolated multiple GNR organisms on the same day were considered one GNR event and classified as polymicrobial bacteremia, except for the purposes of the organism specific analysis.
FQ-resistant GNR bacteremia was defined as the isolation of a GNR organism that was either intermediate or resistant to levofloxacin or ciprofloxacin. Sensitivities to levofloxacin and ciprofloxacin were used because 1) sensitivities against these agents were routinely performed and 2) FQ resistance is known to exhibit a “class effect”, where a decrease in susceptibility to one drug likely means a similar decrease in all FQs.15 FQ sensitivities were determined by the University of Washington Medical Center Microbiology Laboratory using Kirby-Bauer or E-tests, and interpreted using current Clinical and Laboratory Standards Institute (CLSI) guidelines at the time of GNR isolation.16 Bacterial species that did not have current CLSI breakpoints for FQs at the time of specimen collection were excluded from resistance analyses. When evaluating FQ resistance, events were included if a previously isolated organism became resistant on follow-up blood cultures even if those were within 14 days of the initial culture. Similarly, polymicrobial bacteremia events were classified as FQ-resistant event if any of the isolated GNRs were determined to be intermediate or resistant to FQs, except for the purposes of the organism specific analysis, where the sensitivities of each organism were considered separately.
Statistical analysis
For the primary evaluation of GNR incidence, we considered only the first GNR bacteremia event per transplant. The incidence rates of GNR bacteremia and FQ-resistant GNR bacteremia during 30 and 100 days post-transplant were calculated for each calendar year interval. Each patient contributed patient-days at risk from the day of transplant until death, re-transplant, 30 or 100 days post-transplant, or first GNR bacteremia event, whichever occurred first. Changes in incidence rates over time were assessed using a Poisson regression model, with time in one-year intervals as the main independent variable, and count of a GNR bacteremia event post-transplant as the dependent variable. Patient-days (PD) at risk were included as an offset term to account for the varying follow-up time among transplants. Incidence rate ratios (IRR) were used as the measure of change. Clustered robust standard errors were used to account for the correlation between transplants of patients who underwent multiple transplants during our time period of interest.
Next, we conducted analyses considering all GNR bacteremia events for each transplant. In these analyses, each patient contributed patient-days at risk from the day of transplant until death, 30 or 100 days post-transplant, or re-transplant, whichever occurred first. The dependent variable in this analysis was the total count of GNR bacteremia events post-transplant. Changes in incidence rates over time were assessed using similar methods to those described above.
For all analyses, estimated changes in incidence rates were first calculated in unadjusted models to assess the average overall change regardless of the mechanism. We then constructed models that included known risk factors for GNR bacteremia. These covariates were selected a priori and included: age at transplant, severity of underlying illness (low, medium, and high), conditioning regimen score (nonmyeloablative, non-total-body irradiation, total-body irradiation with ≤ 12 Gray (Gy), total-body irradiation with >12 Gy), presence of severe gut graft versus host disease (GVHD, ≥ grade 2), and graft type (bone marrow, peripheral blood, or cord blood). Severity of underlying illnesses categories were defined by outcomes previously observed at our center, while conditioning regimens were first divided into nonmyeloablative and myeloablative, with myeloablative further subdivided by dose of total-body-irradiation used.17 Following the examination of the main results, we elected to conduct post-hoc exploratory analyses to quantify the trends of GNR bacteremia and FQ-resistant GNR bacteremia incidence rates between 2003 and 2009 and separately for 2009 to 2012. These analyses were conducted for first events, all events, day 0–30, and day 0–100. We also evaluated trends in the organism specific incidence rates of FQ-resistant variants of the three most commonly isolated GNR organisms: Escherichia coli, Klebsiella pneumoniae, and Pseudomonas aeruginosa; changes in incidence rates over time were assessed using similar methods.
Lastly, the 30-day all-cause cumulative mortality was compared between patients who developed FQ-resistant GNR bacteremia and those who developed FQ-sensitive GNR bacteremia using Kaplan Meier estimates and the log-rank test. Cox proportional hazards regression models were used to compare the risk of death during 30-days following the first positive GNR blood culture. These analyses were performed both without adjustment for any covariates and with adjustment for underlying disease severity, conditioning regimen score, and age at transplant. All analyses were performed using Stata version 13 (StataCorp, College Station, TX).
RESULTS
Of the 2306 transplants included in this cohort, 280 (12.1%) experienced at least one GNR bacteremia event during the first 100 days post-transplant; 82 of these events occurred between days 0 and 30 and 198 occurred between days 31 and 100. In the first event analysis of patients with available resistance data (n=255), there were 88/255 (34.5%) FQ-resistant events, and 167/255 (65.5%) FQ-sensitive events; 25 GNR events had no resistance data available. The selected demographic and clinical characteristics of patients who experienced a GNR bacteremia event and those that did not were very similar, with the exception that patients who experienced GNR bacteremia were more likely to have severe gut GVHD (Table 1).
Table 1.
Selected characteristics of adult allogeneic HCT transplants by occurrence of GNR bacteremia within 100 days post-transplant
| Variable | GNR event n=280 n (%) |
No GNR events n=2026 n (%) |
|---|---|---|
| Age (years)—median (IQR) | 53 (18) | 51 (20) |
| Sex | ||
| Male | 147 (52.5) | 1216 (60.0) |
| Female | 133 (47.5) | 810 (40.0) |
| Race | ||
| Caucasian | 215 (80.2) | 1654 (85.0) |
| Black | 10 (3.7) | 28 (1.4) |
| Hispanic | 16 (6.0) | 59 (3.0) |
| Asian/Pacific Islander | 13 (4.9) | 100 (5.1) |
| Native American | 2 (0.8) | 18 (0.9) |
| Other | 12 (4.5) | 87 (4.5) |
| Stem-cell source | ||
| Bone marrow | 44 (15.7) | 273 (13.5) |
| Bone marrow and PBSC | 2 (0.7) | 2.9 (0.10) |
| PBSC | 213 (76.1) | 1643 (81.1) |
| Cord blood | 21 (7.5) | 108 (5.3) |
| Diagnosis | ||
| Acute leukemia | 130 (46.4) | 935 (46.2) |
| Multiple myeloma | 16 (5.7) | 114 (5.6) |
| Myelodysplastic syndrome | 43 (15.4) | 402 (19.8) |
| Non-Hodgkin lymphoma | 36 (12.9) | 210 (10.4) |
| Other | 55 (19.6) | 365 (18.0) |
| Underlying Disease Severity | ||
| Low | 38 (13.6) | 287 (14.2) |
| Medium | 143 (51.1) | 1013 (50.0) |
| High | 99 (35.4) | 726 (35.8) |
| Conditioning regimen score | ||
| Nonmyeloablative | 125 (44.6) | 827 (40.8) |
| Non-total-body irradiation | 77 (27.5) | 664 (32.8) |
| Total-body irradiation with ≤ 12 Gy | 67 (23.9) | 471 (23.3) |
| Total-body irradiation with >12 Gy | 11 (3.9) | 64 (3.2) |
| Severe Gut GVHD (≥ grade 2)* | ||
| Yes | 61 (21.8) | 175 (8.7) |
| No | 219 (78.2) | 1845 (91.3) |
Number does not add to n due to missing data
Time trends in incidence rate of all GNR bacteremia
When including only the first GNR bacteremia event per transplant, the overall incidence rate of GNR bacteremia was 1.36 events per 1000 PD (95% CI 1.21, 1.53); the incidence rate between days 0 and 30 was 1.18 per 1000 PD (95% CI 0.94, 1.47) and between days 31 an 100 was 1.46 per 1000 PD (95% CI 1.26, 1.67). Overall incidence rates varied over time, starting at 0.86 events per 1000 PD in 2003 (95%CI 0.52, 1.34), peaking at 2.33 events per 1000 PD in 2009 (95% CI 1.71, 3.11), and declining to its lowest in 2012 at 0.63 events per 1000 PD (95% CI 0.34, 1.08) (Figure 1a). On average, the incidence rate of GNR bacteremia increased annually between 2003 and 2012 by 1%, although this trend was not statistically significant (IRR=1.01, 95%CI: 0.98, 1.05) (Table 2).
Figure 1.
Figure 1a: Incidence rate of first GNR bacteremia event, by transplant year, 2003–2012
Figure 1b: Incidence rate of all GNR bacteremia events, by transplant year, 2003–2012
Table 2.
Unadjusted and adjusted time trends in incidence rates of GNR bacteremia, 2003–2012
|
|
|||||||
|---|---|---|---|---|---|---|---|
| Overall | FQ-resistant | ||||||
|
| |||||||
| 2003–2012 | 2003–2009 | 2009–2012 | 2003–2012 | 2003–2009 | 2009–2012 | ||
| Unadjusted | Unadjusted | ||||||
|
| |||||||
| IRR (95% CI)a | IRR (95% CI)a | ||||||
|
| |||||||
| First Event | Overall (0–100) | 1.01 (0.98, 1.05) | 1.16 (1.09, 1.25) | 0.68 (0.56, 0.80) | 1.01 (0.95, 1.08) | 1.18 (1.04, 1.33) | 0.62 (0.45, 0.86) |
| Month 1 (0–30) | 1.02 (0.95, 1.10) | 1.05 (0.92, 1.20) | 0.82 (0.62, 1.10) | 1.08 (0.96, 1.21) | 1.12 (0.89, 1.40) | 0.87 (0.57, 1.32) | |
|
| |||||||
| Multiple Events | Overall (0–100) | 1.02 (0.98, 1.06) | 1.18 (1.10, 1.26) | 0.69 (0.57, 0.82) | 1.00 (0.94, 1.07) | 1.17 (1.03, 1.33) | 0.62 (0.46, 0.85) |
| Month 1 (0–30) | 1.03 (0.96, 1.11) | 1.08 (0.94, 1.23) | 0.86 (0.64, 1.15) | 1.08 (0.96, 1.21) | 1.12 (0.89, 1.40) | 0.87 (0.57, 1.32) | |
|
| |||||||
| Adjusted | Adjusted | ||||||
|
| |||||||
| aIRR (95% CI)a,b | aIRR (95% CI)a,b | ||||||
|
| |||||||
| First Event | Overall (0–100) | 1.00 (0.97, 1.04) | 1.14 (1.06, 1.23) | 0.67 (0.55, 0.80) | 1.00 (0.93, 1.07) | 1.14 (1.00, 1.30) | 0.62 (0.45, 0.87) |
| Month 1 (0–30) | 1.00 (0.92, 1.08) | 0.97 (0.84, 1.12) | 0.86 (0.63, 1.18) | 1.04 (0.91, 1.18) | 1.00 (0.78, 1.28) | 0.88 (0.55, 1.40) | |
|
| |||||||
| Multiple Events | Overall (0–100) | 1.02 (0.98, 1.05) | 1.16 (1.08, 1.25) | 0.68 (0.56, 0.82) | 0.99 (0.93, 1.06) | 1.13 (1.00, 1.28) | 0.63 (0.46, 0.86) |
| Month 1 (0–30) | 1.01 (0.93, 1.09) | 1.00 (0.87, 1.16) | 0.90 (0.65, 1.25) | 1.04 (0.91, 1.18) | 1.00 (0.78, 1.28) | 0.88 (0.56, 1.38) | |
IRR = incidence rate ratio, CI = confidence interval
Adjusted for underlying disease severity, conditioning regimen score, presence of severe gut GVHD, graft type, and age at transplant.
A post-hoc analysis showed an average annual increase of the incidence rate of GNR bacteremia of 16% (IRR= 1.16, 95%CI: 1.09, 1.25) between 2003 and 2009 and an average annual decrease of 32% (IRR=0.68, 95%CI 0.56, 0.80) between 2009 and 2012. When only considering events that occurred between days 0 and 30 post-transplant, all trends over time were in the same direction as described above, but none were statistically significant (Table 2). Adjusting for known risk factors for bacteremia did not meaningfully change the associations observed in unadjusted analyses (Table 2). Results from the multiple events analyses demonstrated similar results, with minimal increases in the incidence rates of GNR bacteremia compared to the first event analysis (Table 2 and Figure 1b).
Time trends in incidence rate of FQ-resistant GNR bacteremia
The incidence rate of FQ-resistant GNR bacteremia generally displayed similar patterns. In the first event analysis, the overall incidence rate of FQ-resistant GNR bacteremia was 0.43 events per 1000 PD (95% CI 0.34, 0.53). The incidence rates varied over time, starting at 0.23 events per 1000 PD in 2003 (95% CI 0.07, 0.53), peaking at 0.81 events per 1000 PD in 2009 (95% CU 0.46, 1.32), and decreasing to 0.19 events per 1000 PD in 2012 (95% CI 0.05, 0.50) (Figure 1a). On average, the incidence rate of FQ-resistant GNR bacteremia increased annually between 2003 and 2012 by 1%, although this trend was not significant (IRR=1.01, 95%CI: 0.95, 1.08) (Table 2).
A post-hoc analysis showed an average annual increase of the incidence rate of FQ-resistant GNR bacteremia of 18% (IRR= 1.18, 95%CI: 1.04, 1.33) between 2003 and 2009 and an average annual decrease of 38% (IRR=0.62, 95% CI 0.45, 0.86) between 2009 and 2012. When only considering events that occurred between days 0 and 30 post-transplant, all trends over time were in the same direction as described above, but none were statistically significant (Table 2). Adjusting for known risk factors for bacteremia did not meaningfully change the associations observed in unadjusted analyses (Table 2). Results from the multiple events analysis again found similar results, including minimal overall increases in the incidence rates of FQ-resistant GNR bacteremia compared to the first event analysis (Table 2 and Figure 1b). The three most commonly isolated organisms were Escherichia coli, Klebsiella pneumoniae, and Pseudomonas aeruginosa, which demonstrated overall FQ-sensitivity of 27.8%, 62.5%, and 58.1%, respectively (Table 3); individual analyses of annual trends in the incidence rates of FQ-resistant variants of these three organisms showed no significant annual trends between 2003 and 2012 (IRR=1.08, 95% CI 0.96, 1.20; IRR=1.09 95% CI 0.99, 1.19; IRR=0.91 95%CI 0.77, 1.08, respectively).
Table 3.
Most common gram-negative rods isolated from blood cultures and their respective fluoroquinolone sensitivities by year, 2003–2012
| All | 2003 | 2004 | 2005 | 2006 | 2007 | 2008 | 2009 | 2010 | 2011 | 2012 | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||||||||||||||
| n | %Sa | n | %S | n | %S | n | %S | n | %S | n | %S | n | %S | n | %S | n | %S | n | %S | n | %S | |
| E. coli | 54 | 27.8 | 1 | 100 | 6 | 16.7 | 2 | 50 | 9 | 11.1 | 10 | 50.0 | 2 | 50.0 | 7 | 28.6 | 8 | 37.5 | 6 | 0 | 3 | 0 |
| K. pneumoniae | 48 | 62.5 | 0 | -- | 1 | 100 | 5 | 60 | 10 | 90.0 | 3 | 66.7 | 9 | 44.4 | 9 | 22.2 | 6 | 66.7 | 2 | 100 | 3 | 100 |
| P. aeruginosa | 43 | 58.1 | 2 | 0 | 7 | 71.4 | 8 | 37.5 | 3 | 66.7 | 2 | 50 | 6 | 93.3 | 7 | 57.1 | 3 | 66.7 | 2 | 50 | 3 | 66.7 |
| S. maltophilia | 40 | 67.5 | 2 | 0 | 1 | 0 | 4 | 50 | 3 | 100 | 9 | 55.6 | 5 | 60.0 | 6 | 83.3 | 2 | 100 | 7 | 100 | 1 | 0 |
| K. oxytoca | 34 | 85.3 | 2 | 50 | 2 | 100 | 7 | 85.7 | 2 | 100 | 4 | 100 | 3 | 100 | 8 | 87.5 | 1 | 100 | 5 | 60 | 0 | -- |
| E. cloacae | 27 | 85.2 | 2 | 100 | 1 | 100 | 4 | 100 | 3 | 100 | 4 | 100 | 7 | 71.4 | 6 | 66.7 | 0 | -- | 0 | -- | 1 | 100 |
Percent sensitive to fluoroquinolones
Survival analysis by FQ resistance
Patients who had an initial FQ-resistant GNR bacteremia event had a significantly higher 30-day post-event cumulative mortality than patients who experienced a FQ-sensitive event (20.8% vs. 9.6%, p=0.018) (Figure 2). In an unadjusted survival analysis, patients with FQ-resistant GNR bacteremia had an increased risk of death through 30-days post GNR isolation than patients with FQ-sensitive bacteremia (HR 2.23, 95% CI: 1.13, 4.43). This association persisted after adjustment for severity of underlying illness, conditioning regimen score, and age at transplant (HR 2.11, 95% CI: 1.06, 4.23).
Figure 2.
Kaplan-Meier survival estimates 30-days post first infection, by FQ resistance status
DISCUSSION
In this large single center longitudinal retrospective study, we examined trends in the incidence of GNR bacteremia and FQ-resistant GNR bacteremia among adult allogeneic HCT recipients over a decade during which levofloxacin was used for neutropenic prophylaxis. We found that, on average, there was no significant trend in the incidence rates of GNR bacteremia and FQ-resistant GNR bacteremia between 2003 and 2012 in this population. In post-hoc analyses, we found that the incidence rates of GNR bacteremia and FQ-resistant GNR bacteremia increased annually between 2003 and 2009, and then decreased between 2009 and 2012. An organism specific analysis did not reveal any significant trends in the incidence rates of FQ-resistant Escherichia coli, Klebsiella pneumoniae, or Pseudomonas aeuroginosa, the three most common organisms in this cohort. Lastly, these data demonstrate that patients who developed bacteremia from FQ-resistant GNRs had higher 30-day mortality than those who developed bacteremia from FQ-sensitive GNRs.
Other centers have described rising rates of GNR infections in HCT patients, but few studies have studied trends in GNR bacteremia over time, or assessed such data after a major change in neutropenic antibiotic prophylaxis. The findings of studies that quantified this association have been inconsistent, with at least one center reporting significantly increasing rates of GNR bacteremia during use of levofloxacin prophylaxis,1 and another reporting no significant change.10 Available data on the issue of FQ-resistance in patient populations receiving FQ prophylaxis also varies between centers. Some have described non-significant increases in the proportion of GNR isolates that are FQ-resistant during an era of FQ prophylaxis,1,3 while others, including a previous study at our center, reported no changes in FQ resistance after initiation of levofloxacin prophylaxis.9,12 In contrast, other centers have described significantly increasing rates of FQ-resistant GNR during FQ prophylaxis,10,18,19 including one that only measured FQ-resistant Escherichia coli.18 Our data are most consistent with studies that identified no significant overall increase in GNR bacteremia or FQ-resistant GNR bacteremia over time. To our knowledge, none of these studies have addressed survival differences between HCT patients who developed bacteremia with FQ-resistant and FQ-sensitive GNRs, however, FQ resistance has been identified as an independent risk factor for death after infection in other populations.20
The observed patterns of incidence rates in this study were somewhat unexpected, since prior data from our center indicated that rates of GNR bacteremia had been increasing.11 We hypothesized that this increase might be associated with our widespread use of levofloxacin prophylaxis, especially if the increasing rates of GNR bacteremia were accompanied by increasing rates of FQ-resistant GNRs. Our post-hoc analyses confirmed this increase through 2009, but rates of GNR bacteremia steeply decreased between 2009 and 2012. There are likely several contributing factors to the decrease in GNR bacteremia events after 2009 in this population. Several infection control interventions were implemented between 2009 and 2010, including the initiation of chlorhexidine gluconate (CHG) baths in January 2010, the development and implementation of a new line bundle also in January 2010, adoption of “scrub the hub” 21 as a standardized protocol in September 2010, and a switch to CHG impregnated dressings in February 2009. Since such interventions have been associated with decreased rates of bloodborne infections,22–24 it is possible that together, they contributed to the decrease in GNR bacteremia between 2009 and 2012. Temporal changes in GVHD incidence and therapy, such as alterations in glucocorticoid use for initial GVHD treatment,25 may also have contributed to the observed decrease in GNR bacteremia. Of note, the post-2009 decline in the incidence of GNR bacteremia was not appreciably affected by adjusting for conditioning regimen (Table 2), suggesting that the decline could not be attributed to a change in the number of patients receiving reduced-intensity conditioning.
It is also important to acknowledge that standard practice changed in 2010 to discontinue collection of an extra set of blood cultures that were held for yeast and fungi. With improvements in microbiologic techniques, such methods provided no additional benefit in isolating fungal pathogens in these patients. It is possible that these extra cultures cultivated bacterial growth, and their discontinuation could have contributed to the decrease in isolation of GNR organisms. Additionally, the microbiology changed their blood culture system in 2010, potentially resulting in differential isolation of organisms. To address these issues, we examined the number of blood cultures ordered between 2003 and 2012, and these data demonstrated a similar time trend as the incidence rates of GNR bacteremia, suggesting that these laboratory changes did not contribute significantly to our results (data not shown). While we hypothesize that many of the aforementioned factors influenced the observed recent decline in rates of GNR bacteremia, this study was not designed to directly attribute changes in incidence to any of these infection control interventions or laboratory variations.
Most importantly, we did not observe an increase in FQ-resistant GNR bacteremia with widespread use of levofloxacin as neutropenic prophylaxis in this population. One reason for this could be shorter exposure intervals to levofloxacin in the overall population driven by a gradual increase in the number of nonmyeloablative transplants over time. Another possible explanation is that our patients receive a higher dose of levofloxacin for prophylaxis (750 mg) than that tested in placebo controlled trials (500 mg).26,27 FQs are unique in that they are synthetic antibiotics, and although it was thought that they might be more insulated from resistance issues, overuse in medicine and in agriculture has led to reports of increasing rates of FQ resistance worldwide.28 Lack of a significant increase in FQ-resistance observed in this study suggests that levofloxacin may continue to be a viable neutropenic prophylaxis agent in this population.
Lastly, these data show that patients with FQ-resistant GNR bacteremia had a greater than two-fold increased risk of death within 30-days post-infection compared to patients with FQ-sensitive GNR bacteremia. This association persisted even after adjusting for underlying disease severity, type of conditioning regimen, and age at transplant, suggesting that FQ resistance may be an independent risk factor for death in patients with GNR bacteremia. One mechanism that may explain such differences is that patients with resistant infections experience a delay in receiving adequate antimicrobial therapy. Previous studies have demonstrated an association between administration of inadequate antimicrobial treatment and mortality.29,30 Another mechanism might be that organisms that demonstrate lower rates of FQ-sensitivity, such as Pseudomonas aeruginosa (Table 3), have been shown to have associated increased mortality.31 Further research is needed to elucidate risk factors for the development of FQ-resistant infections, to allow for prompt identification and treatment of these patients. Overall, these data highlight the serious nature of antimicrobial resistance and the importance of continued monitoring of FQ resistance trends in HCT recipients and other high-risk populations.
The retrospective and observational nature of this study imposes limits on the interpretation of our data. There are likely variables that have changed over time and that influence GNR bacteremia rates or FQ resistance for which we could not adjust. One specific unmeasured variable was the proportion of the population that did not receive any early post-transplant levofloxacin. However, we estimate that this subset of the population would make up less than 5% of this cohort, and we do not feel that minor changes in that proportion over time would appreciably change our results. Additionally, there were organisms that lacked FQ sensitivity data and could not be included in the FQ resistance analysis, and it is possible that changes in the incidence of these unclassifiable organisms over time may have had a minimal effect on our results. Finally, these data only reflect the experience of a single center and may not be generalizable to other institutions, as regional variances in transplant conditioning, antimicrobial therapy, and prevalence of FQ-resistance could impact GNR bacteremia trends at other centers. Strengths of this study include the large sample size and the valuable long-term longitudinal data, which inform evidence based decisions about use of FQ prophylaxis at our center.
In summary, these data demonstrate that rates of FQ-resistant GNR bacteremia have not significantly increased during an era of levofloxacin prophylaxis in adult allogeneic HCT recipients at our large comprehensive cancer. Recent decreases in incidence rates of GNR bacteremia were potentially a result of a combination of center-wide changes, including several infection control interventions. Although there is no evidence that levofloxacin prophylaxis is associated with an increase in FQ-resistant bacteremia in HCT recipients at our center, the increased mortality associated with FQ-resistant GNR bacteremia re-enforces the importance of monitoring emerging FQ resistance in this high-risk population.
GNR in transplant highlights.
We examined GNR bacteremia during 10 years of FQ prophylaxis in HCT recipients
Incidence of FQ-resistant GNR bacteremia did not significantly change over time
Bacteremia caused by FQ-resistant GNRs were associated with higher mortality
Centers should monitor rates of FQ resistance in populations with high FQ exposure
Acknowledgments
Support: S.A.P. is supported by NIH Grant K23HL096831.
Footnotes
Presentation of material in submitted manuscript: Data from this manuscript were presented in part at the 54th Annual Interscience Conference on Antimicrobial Agents and Chemotherapy in Washington DC, September 2014
Disclosure of Conflicts of Interest: S.A.P. has received research support and been a consultant for Merck, Sharp and Dohme, Corp. and Optimer/Cubist Pharmaceuticals. All other authors report no conflicts.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Mikulska M, Del Bono V, Raiola AM, et al. Blood stream infections in allogeneic hematopoietic stem cell transplant recipients: reemergence of Gram-negative rods and increasing antibiotic resistance. Biol Blood Marrow Transplant. 2009;15(1):47–53. doi: 10.1016/j.bbmt.2008.10.024. [DOI] [PubMed] [Google Scholar]
- 2.Collin Ba, Leather HL, Wingard JR, Ramphal R. Evolution, incidence, and susceptibility of bacterial bloodstream isolates from 519 bone marrow transplant patients. Clin Infect Dis. 2001;33(7):947–53. doi: 10.1086/322604. [DOI] [PubMed] [Google Scholar]
- 3.Bock AM, Cao Q, Ferrieri P, Young J-AH, Weisdorf DJ. Bacteremia in blood or marrow transplantation patients: clinical risk factors for infection and emerging antibiotic resistance. Biol Blood Marrow Transplant. 2013;19(1):102–8. doi: 10.1016/j.bbmt.2012.08.016. [DOI] [PubMed] [Google Scholar]
- 4.Gooley TA, Chien JW, Pergam SA, et al. Reduced mortality after allogeneic hematopoietic-cell transplantation. N Engl J Med. 2010;363(22):2091–101. doi: 10.1056/NEJMoa1004383. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Tomblyn M, Chiller T, Einsele H, et al. Guidelines for preventing infectious complications among hematopoietic cell transplantation recipients: a global perspective. Biol Blood Marrow Transplant. 2009;15(10):1143–238. doi: 10.1016/j.bbmt.2009.06.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.National Comprehensive Cancer Network. [October 2014];Prevention and treatment of cancer related infections. Version 2.2014. Accessed at www.nccn.org on.
- 7.Girmenia C, Menichetti F. Neutropenia and Infections Current Epidemiology and Prevention of Infectious Complications. 2011. [Google Scholar]
- 8.Gafter-Gvili A, Fraser A, Paul M, et al. Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy. Cochrane database Syst Rev. 2012;1(4):CD004386. doi: 10.1002/14651858.CD004386.pub3. [DOI] [PubMed] [Google Scholar]
- 9.Guthrie Ka, Yong M, Frieze D, Corey L, Fredricks DN. The impact of a change in antibacterial prophylaxis from ceftazidime to levofloxacin in allogeneic hematopoietic cell transplantation. Bone Marrow Transplant. 2010;45(4):675–81. doi: 10.1038/bmt.2009.216. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Therriault BL, Wilson JW, Barreto JN, Estes LL. Characterization of bacterial infections in allogeneic hematopoietic stem cell transplant recipients who received prophylactic levofloxacin with either penicillin or doxycycline. Mayo Clin Proc. 2010;85(8):711–8. doi: 10.4065/mcp.2010.0006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Green ML, Leisenring W, Stachel D, et al. Efficacy of a viral load-based, risk-adapted, preemptive treatment strategy for prevention of cytomegalovirus disease after hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2012;18(11):1687–99. doi: 10.1016/j.bbmt.2012.05.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Simondsen Ka, Reed MP, Mably MS, Zhang Y, Longo WL. Retrospective analysis of fluoroquinolone prophylaxis in patients undergoing allogeneic hematopoietic stem cell transplantation. J Oncol Pharm Pract. 2013;19(4):291–7. doi: 10.1177/1078155212465215. [DOI] [PubMed] [Google Scholar]
- 13.Lautenbach E, Metlay JP, Bilker WB, Edelstein PH, Fishman NO. Association between fluoroquinolone resistance and mortality in Escherichia coli and Klebsiella pneumoniae infections: the role of inadequate empirical antimicrobial therapy. Clin Infect Dis. 2005;41(7):923–9. doi: 10.1086/432940. [DOI] [PubMed] [Google Scholar]
- 14.Tatro JB. Endogenous antipyretics. Clin Infect Dis. 2000;31(Suppl 5):S190–201. doi: 10.1086/317519. [DOI] [PubMed] [Google Scholar]
- 15.Becnel Boyd L, Maynard MJ, Morgan-Linnell SK, et al. Relationships among ciprofloxacin, gatifloxacin, levofloxacin, and norfloxacin MICs for fluoroquinolone-resistant Escherichia coli clinical isolates. Antimicrob Agents Chemother. 2009;53(1):229–34. doi: 10.1128/AAC.00722-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.CLSI document M100-S22. 2012. Performance Standards for Antimicrobial Susceptibility Testing: Twenty-second International Supplement. [Google Scholar]
- 17.Parimon T, Au DH, Martin PJ, Chien JW. A risk score for mortality after allogeneic hematopoietic cell transplantation. [Accessed April 18, 2014];Ann Intern Med. 2006 144(6):407–14. doi: 10.7326/0003-4819-144-6-200603210-00007. Available at: http://www.ncbi.nlm.nih.gov/pubmed/16549853. [DOI] [PubMed] [Google Scholar]
- 18.Kern WV, Klose K, Jellen-Rittera S, et al. Fluoroquinolone resistance of Escherichia coli at a cancer center: epidemiologic evolution and effects of discontinuing prophylactic fluoroquinolone use in neutropenic patients with leukemia. Eur J Clin Microbiol Infect Dis. 2005;24(2):111–8. doi: 10.1007/s10096-005-1278-x. [DOI] [PubMed] [Google Scholar]
- 19.Macesic N, Morrissey CO, Cheng aC, Spencer a, Peleg aY. Changing microbial epidemiology in hematopoietic stem cell transplant recipients: increasing resistance over a 9-year period. Transpl Infect Dis. 2014;(5):1–10. doi: 10.1111/tid.12298. [DOI] [PubMed] [Google Scholar]
- 20.Lautenbach E, Metlay JP, Bilker WB, Edelstein PH, Fishman NO. Association between fluoroquinolone resistance and mortality in Escherichia coli and Klebsiella pneumoniae infections: the role of inadequate empirical antimicrobial therapy. Clin Infect Dis. 2005;41(7):923–9. doi: 10.1086/432940. [DOI] [PubMed] [Google Scholar]
- 21.Simmons S, Bryson C, Porter S. “Scrub the hub”: cleaning duration and reduction in bacterial load on central venous catheters. Crit Care Nurs Q. 2011;34(1):31–5. doi: 10.1097/CNQ.0b013e3182048073. [DOI] [PubMed] [Google Scholar]
- 22.Bleasdale SC, Trick WE, Gonzalez IM, Lyles RD, Hayden MK, Weinstein Ra. Effectiveness of chlorhexidine bathing to reduce catheter-associated bloodstream infections in medical intensive care unit patients. Arch Intern Med. 2007;167(19):2073–9. doi: 10.1001/archinte.167.19.2073. [DOI] [PubMed] [Google Scholar]
- 23.Galpern D, Guerrero A, Tu A, Fahoum B, Wise L. Effectiveness of a central line bundle campaign on line-associated infections in the intensive care unit. Surgery. 2008;144(4):492–5. doi: 10.1016/j.surg.2008.06.004. discussion 495. [DOI] [PubMed] [Google Scholar]
- 24.Sannoh S, Clones B, Munoz J, Montecalvo M, Parvez B. A multimodal approach to central venous catheter hub care can decrease catheter-related bloodstream infection. Am J Infect Control. 2010;38(6):424–9. doi: 10.1016/j.ajic.2009.07.014. [DOI] [PubMed] [Google Scholar]
- 25.Mielcarek M, Furlong T, Storer BE, Green ML, Carpenter PA, McDonald GB, Flowers MED, Storb R, Appelbaum FR, Boeckh M, Martin PJMP. Efficacy and Safety Of Lower-Dose Glucocorticoids For Initial Treatment Of Acute Graft-Versus-Host Disease: A Randomized Controlled Trial. Mielcarek M, editor. Blood. 2013;122(21):703. Available at: http://www.bloodjournal.org/content/122/21/703.abstract.
- 26.Bucaneve G, Micozzi A, Menichetti F, et al. Levofloxacin to prevent bacterial infection in patients with cancer and neutropenia. N Engl J Med. 2005;353(10):977–87. doi: 10.1056/NEJMoa044097. [DOI] [PubMed] [Google Scholar]
- 27.Cullen M, Steven N, Billingham L, et al. Antibacterial prophylaxis after chemotherapy for solid tumors and lymphomas. N Engl J Med. 2005;353(10):988–98. doi: 10.1056/NEJMoa050078. [DOI] [PubMed] [Google Scholar]
- 28.Robicsek A, Jacoby Ga, Hooper DC. The worldwide emergence of plasmid-mediated quinolone resistance. Lancet Infect Dis. 2006;6(10):629–40. doi: 10.1016/S1473-3099(06)70599-0. [DOI] [PubMed] [Google Scholar]
- 29.Ibrahim EH, Sherman G, Ward S, Fraser VJ, Kollef MH. The influence of inadequate antimicrobial treatment of bloodstream infections on patient outcomes in the ICU setting. Chest. 2000;118(1):146–55. doi: 10.1378/chest.118.1.146. Available at: http://www.ncbi.nlm.nih.gov/pubmed/10893372. [DOI] [PubMed] [Google Scholar]
- 30.Leibovici L, Shraga I, Drucker M, Konigsberger H, Samra Z, Pitlik SD. The benefit of appropriate empirical antibiotic treatment in patients with bloodstream infection. J Intern Med. 1998;244(5):379–86. doi: 10.1046/j.1365-2796.1998.00379.x. Available at: http://www.ncbi.nlm.nih.gov/pubmed/9845853. [DOI] [PubMed] [Google Scholar]
- 31.Hakki M, Limaye aP, Kim HW, Kirby Ka, Corey L, Boeckh M. Invasive Pseudomonas aeruginosa infections: high rate of recurrence and mortality after hematopoietic cell transplantation. Bone Marrow Transplant. 2007;39(11):687–93. doi: 10.1038/sj.bmt.1705653. [DOI] [PubMed] [Google Scholar]