- Email: [email protected]
Antifungal use in immunocompetent, critically ill patients with pneumonia does not improve clinical outcomes
Heart & Lung xxx (2016) 1e6
Contents lists available at ScienceDirect
Heart & Lung journal homepage: www.heartandlung.org
Antifungal use in immunocompetent, critically ill patients with pneumonia does not improve clinical outcomes Meagan Griffin, PharmD a, Desiree E. Kosmisky, PharmD c, Megan A. Templin, MPH, MS c, Toan Huynh, MD b, Lewis H. McCurdy III, MD b, Timothy R. Pasquale, PharmD b, Kelly E. Martin, PharmD b, * a
Memorial Sloan Kettering Cancer Center, 1275 York Avenue, New York, NY 10065, USA Carolinas Medical Center, 1000 Blythe Boulevard, Charlotte, NC 28203, USA c Carolinas HealthCare System, Charlotte, NC 28203, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 18 April 2016 Received in revised form 1 August 2016 Accepted 2 August 2016 Available online xxx
Purpose: To determine if treating bronchoalveolar lavage (BAL) culture-positive patients with antifungal therapy impacted mortality compared to not treating due to presumed colonization. Methods: We conducted a retrospective study of immunocompetent, critically ill adult patients from 2010 to 2014. Patients with a BAL culture-positive for Candida or unspeciated yeast and a clinical suspicion of pneumonia were included. The treatment group received an antifungal agent for at least 5 days, and the control group received either no antifungal therapy or an antifungal agent for less than 48 h. Recruitment occurred in a 2:1 ratio of untreated versus treated patients. Results: Seventy-five patients were included. In-hospital mortality was similar between treated and untreated groups (24% vs. 26%, P ¼ 0.85). Length of stay and duration of mechanical ventilation also did not differ between the two groups. Conclusion: We did not observe a difference in mortality or clinical outcomes in patients treated with antifungal agents. Presumptive antifungal therapy for BAL-positive Candida or yeast in immunocompetent patients did not result in improved clinical outcomes. Ó 2016 Elsevier Inc. All rights reserved.
Keywords: Candida Yeast Bronchoscopy Antifungal agents Pneumonia
Introduction Candida species account for 9% of hospital-acquired bloodstream infections and candidemia is associated with a crude mortality of up to 47%.1 However, Candida albicans, the most common species of yeast, is an oropharyngeal colonizer in up to 20% of healthy individuals and 55% of hospitalized patients.2 The diagnosis of candidal pneumonia in non-neutropenic patients is extremely rare. Indeed, candidal pneumonia requires histopathologic confirmation with findings of yeast cells or pseudohyphae on lung biopsy, which is a practice not routinely performed.3 Candida isolated from
Conflict of interest: The authors whose names are listed immediately below certify that they have no disclosures in the subject matter or materials discussed in this manuscript. Meagan Griffin, Desiree E. Kosmisky, Megan A. Templin, Toan Huynh, Kelly E. Martin. The authors whose names are listed immediately below are part of the Cubist Speaker’s Bureau. Timothy R. Pasquale, Lewis H. McCurdy III. * Corresponding author. E-mail address: [email protected] (K.E. Martin). 0147-9563/$ e see front matter Ó 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.hrtlng.2016.08.002
bronchoscopic specimens is usually thought to be non-pathogenic.2 Several studies have confirmed that isolation of Candida from a bronchoalveolar lavage (BAL), although common in hospitalized patients, is rarely an active source of pneumonia.3e5 As such, current guidelines for candidiasis do not recommend the initiation of antifungal treatment based on a positive BAL alone due to the frequent rate of colonization, the poor predictive value of this finding, and the rarity of candidal pneumonia.6 Despite these guideline recommendations, providers frequently prescribe antifungal agents for critically ill, febrile patients with no clear evidence of invasive fungal infection. Indeed, our recent unpublished institutional survey of 64 providers from trauma surgical critical care, pulmonary critical care, and infectious diseases demonstrated significant variability in self-reported prescribing practices of antifungal therapy in critically ill patients. Up to 24% of respondents reported the empiric use of antifungal therapy for clinical suspicion of pneumonia, and 42% would initiate antifungal treatment of BAL cultures positive for C. albicans. Overall, trauma surgical critical care providers were most likely to consider Candida to be a cause of pneumonia and to prescribe antifungals. By contrast, infectious diseases physicians were the least likely to
2
M. Griffin et al. / Heart & Lung xxx (2016) 1e6
prescribe antifungal therapy for pneumonia. The practice of prescribing antifungals in this clinical setting is not limited to our institution. Our survey results are consistent with a study assessing antifungal prescribing practices in critically ill, non-neutropenic patients in France, where 24% of intensivists reported prescribing antifungal therapy for patients with chronic obstructive pulmonary disease requiring mechanical ventilation with a tracheal aspirate positive for Candida species.7 The overuse of antifungal therapy can lead to unintended consequences including adverse effects and increased resistance.8 Previous azole exposure is a risk factor for subsequent Candida infections with decreased susceptibility profiles, such as Candida glabrata and Candida krusei.9 The Infectious Diseases Society of America (IDSA) guidelines recommend the empiric use of an echinocandin in patients with recent treatment with an azole for this very reason.6 The purpose of the study was to determine if treating patients with antifungal therapy for a BAL culture positive for Candida or unspeciated yeast had an impact on mortality compared to not treating due to presumed colonization. The primary outcome was a comparison of in-hospital mortality between the treatment and non-treatment groups. Secondary outcomes included hospital length of stay (LOS), intensive care unit (ICU) LOS, duration of mechanical ventilation, and time to mortality from BAL culture. Methods Design This was a single-center, retrospective chart review of critically ill adult patients admitted to one of three ICUs (medical ICU, surgical trauma ICU and neurosurgical ICU) at an academic medical center from January 1, 2010 to July 31, 2014. This study received Institutional Review Board approval. Inclusion criteria consisted of age 18 years old, mechanical ventilation at the time of BAL and at least one BAL positive for Candida or unspeciated yeast during ICU admission. Patients who met criteria for clinical suspicion of nosocomial pneumonia on the day of BAL were included. These criteria consisted of the presence of infiltrate on chest radiograph or chest computerized tomography (CT) based on radiology reports, new onset purulent sputum, changes in secretions or secretions noted on BAL, and either a temperature of 38 C or 36 C and/or a white blood cell (WBC) of 12,000 or 4000 cells/mm3. Patients were excluded if they were severely immunocompromised, defined as patients with a history of hematopoietic-stem cell transplantation, chemotherapy in the last 30 days, solid organ transplantation, human immunodeficiency virus (HIV) with a CD4 T-lymphocyte count of <200 cells/mm3, or daily corticosteroid therapy with a dose equivalent to 20 mg of prednisone for 14 days. Patients who received antifungals for reasons other than pneumonia were excluded using the following criteria: medical record documentation of antifungal use for other indications, presence of positive blood, fluid, or urine cultures with Candida species 7 days prior to or during antifungal therapy, receiving an antifungal agent other than fluconazole or an echinocandin, and receiving antifungal treatment duration of >48 hours but <5 days. Patients were also excluded if they were diagnosed with community-acquired pneumonia (CAP), pregnant, had a BAL growing a carbapenem-resistant or pan-resistant organism, previously included in the study, or if there was an absence of data to assess clinical suspicion of pneumonia. Patients with bacteria isolated in a BAL culture could be included if the organism was adequately treated with an antibiotic the organism was susceptible to for at least seven days.
Variables and outcome measures Data included age, gender, race, weight, and comorbidities. Characteristics related to pneumonia consisted of presumed pneumonia, radiographic evidence of pneumonia by either chest radiograph or CT scan, and ratio of partial pressure of arterial oxygen and fraction of inspired oxygen (PaO2/FiO2). Disease severity scores were assessed by Acute Physiology and Chronic Health Evaluation (APACHE) II on day of ICU admission and Clinical Pulmonary Infection Score (CPIS) on day of BAL. Microbiologic data included bacterial species isolated from the respiratory tract and bronchoscopy findings, including purulent sputum. Cultures within 28 days were also reviewed for presence of fluconazole-resistant, or non-albicans Candida species and Pseudomonas species. WBC count and maximum temperature were collected on the day of BAL and persistence of these factors were assessed on day 8 from BAL. Maximum heart rate, respiratory rate or partial pressure of arterial carbon dioxide (PaCO2), systolic blood pressure, and lactate were collected on the day of BAL. Pertinent antimicrobial therapy included duration of concurrent antibiotics, duration of antifungal therapy, time from bronchoscopy to initiation of antifungal therapy, antifungal agent prescribed (fluconazole or echinocandin), fluconazole dose and route, and echinocandin doses. Clinical endpoints consisted of in-hospital mortality, time to mortality from BAL culture, hospital length of stay (LOS), ICU LOS, duration of mechanical ventilation, and cause of death per death records documented within patient charts. Study data were collected and managed using REDCap (Research Electronic Data Capture) tools hosted at Carolinas HealthCare System.10 REDCap is a secure, web-based application which provides an interface for validated data entry, audit trails for tracking data manipulation and export procedures, and automated export procedures for data downloads. Statistical analysis Data were analyzed using descriptive statistics consisting of means, medians, and percentages. The primary analysis was a chisquare test comparing in-hospital mortality for the two treatment groups and the corresponding 95% confidence intervals on the difference in the proportions. Secondary outcomes were analyzed with the non-parametric Wilcoxon rank-sum test. SAS Enterprise GuideÒ, version 5.1. A two-tailed P-value of less than 0.05 was considered statistically significant. Results Due to low enrollment in the treatment group, the IRB-approved protocol was modified to include patients in a 2 to 1 ratio in reverse chronological order until a total of 75 patients were included. There were a total of 25 patients in the treated group and 50 patients in the non-treated group (Fig. 1). A total of 224 patients were excluded from the study, of which 22.8% were found to have evidence of Candida or unspeciated yeast in other cultures. The majority of patients were Caucasian males, with a median age of 59 years old (Table 1). Most patients had clinical suspicion of ventilatorassociated pneumonia (VAP) with a median APACHE II score of 25 in the treated group and 27 in the non-treated group. Of note, APACHE II scores were not able to be calculated for four patients in the treated group and three patients in the non-treated group due to lack of data. Differences between groups occurred in relation to ICU-specialty, with significantly more treated patients residing in the surgical trauma ICU compared to the other ICUs (P ¼ 0.0013). A subgroup analysis of all patients in the medical ICU and surgical trauma ICU showed average APACHE II scores of 34 and 29 respectively (P ¼ 0.2684). Additionally, mortality between medical
M. Griffin et al. / Heart & Lung xxx (2016) 1e6
3
316 posiƟve paƟent bronchoscopies screened
224 Excluded: 62 Previously included paƟents 51 Candida/yeast other cultures 30 No infiltrate on CXR/CT 25 Immunocompromised 13 Diagnosed CAP 11 AnƟfungal therapy > 48 hours but < 5 days 10 Didn't meet clinical criteria 8 Not mechanically venƟlated
17 non-treated paƟents excluded when goal inclusion of 50 non-treated paƟents met
5 Already on anƟfungal agent 4 Bacterial pathogens not treated properly 3 Absence of data to assess pneumonia 1 Age < 18 years old 1 Use of other anƟfungal agent
Treated group:
Non-treated group:
25 paƟents
50 paƟents
Fig. 1. Patient enrollment.
ICU and surgical trauma ICU patients was found to be similar (37.5% vs 17.14%, P ¼ 0.0604). Overall, no statistically significant difference was observed between the treated and non-treated groups for the primary endpoint of in-hospital mortality (24% vs. 26%, P ¼ 0.8511), or the secondary endpoints of time to mortality from BAL culture, pneumonia clinical cure at day 8 from BAL, hospital LOS, ICU LOS, and duration of mechanical ventilation (Table 2). Additionally, the median number of ventilator-free days was 4 in the treated group and 6 in the nontreated group (P ¼ 0.4766). C. albicans accounted for the majority of Candida isolates in both groups (Table 3). Additionally, unspeciated yeast was isolated in 20% of treated and 32% of non-treated patients. There were a total of 26 patients who had a repeat BAL within 28 days; Pseudomonas was isolated in three patients. There was no incidence of fluconazole-resistant Candida or non-albicans Candida on any repeat respiratory culture. Forty-eight percent of patients received fluconazole for treatment of their presumed Candida pneumonia, 36% received caspofungin and 16% switched between both agents (Table 4). Of the patients who received fluconazole, 93.8% received intravenous therapy. Seventy-five percent received a dose of 400 mg daily and 25% received a dose of 200 mg daily. A total of 18.8% of patients received a fluconazole loading dose that was twice the maintenance dose. Of the patients who received caspofungin,
76.9% were treated with a 70 mg loading dose followed by 50 mg daily, 15.4% received 50 mg daily without a loading dose and 7.7% of patients received a 70 mg loading dose followed by 35 mg daily to adjust for hepatic insufficiency. Discussion To our knowledge, this is the first study conducted to focus on treatment-related outcomes of patients with Candida-positive BALs. Our results demonstrated that treatment of Candida-positive BALs with antifungal therapy had no significant impact on inhospital mortality, time to mortality, hospital LOS, or duration of mechanical ventilation compared to patients who were not treated. Several previous studies have investigated the impact of Candida colonization in critically ill patients. Schnabel and colleagues examined whether Candida pneumonia was a clinical concern within the critically ill population.11 Of 701 BAL samples analyzed, only 0.7% correlated with a proven candidal pneumonia. These findings suggested that true fungal pneumonia in this setting is rare despite frequent colonization.11 In a retrospective analysis of patients with diagnosed VAP conducted by Delisle and colleagues, colonization of Candida within the respiratory tract was associated with increased hospital mortality and increased hospital LOS.12 However, neither of these studies were focused on
4
M. Griffin et al. / Heart & Lung xxx (2016) 1e6
Table 1 Baseline characteristics of patients with Candida-positive bronchoscopy cultures and clinical suspicion of pneumonia.
Table 3 Microbial information of patients with Candida-positive bronchoscopy cultures and clinical suspicion of pneumonia.
Characteristics
Treated (N ¼ 25)
Not treated (N ¼ 50)
P value
Microbial information
Age, years, median SD Weight, kg, median SD Gender, male, n (%) Race American Indian or Alaskan Native Black or African American Native Hawaiian or Pacific Islander White Unknown Comorbid conditions, n (%) Diabetes COPD Renal failure Malignancy Liver disease Post-surgical Intensive care unit (ICU), n (%) Medical ICU Surgical trauma ICU Neurosurgical ICU Year of admission, n (%) 2011 2012 2013 2014 SIRS criteria met Severe sepsis criteria met CPIS suggestive of pneumonia, n (%) Yes No Unable to determine Type of pneumonia, n (%) Ventilator-associated Hospital-acquired Healthcare-associated APACHE II score, median SD
60 16.5 90 21.7 18 (72.0)
59 15.5 82 27.8 30 (60.0)
0.9140 0.9418 0.3074 0.7521
1 (4.0)
0
3 (12.0) 0
7 (14.0) 1 (2.0)
20 (80.0) 1 (4.0)
39 (78.0) 3 (6.0)
6 7 6 6 3 9
(24.0) (28.0) (24.0) (24.0) (12.0) (36.0)
13 (26.0) 12 (24.0) 10 (20.0) 6 (12.0) 7 (14.0) 12 (24.0)
Type of Candida isolated, n (%) Candida albicans Candida tropicalis Bacteria isolated from BAL, n (%) Type of bacteria isolated, n (%) Pseudomonas aeruginosa Klebsiella spp. Haemophilus influenzae Acinetobacter spp. Streptococcus pneumoniae MSSA MRSA Other Antibiotics started, n (%) Duration of antibiotics, days, median SD
5 (20.0) 19 (76.0) 1 (4.0)
27 (54.0) 16 (32.0) 7 (14.0)
10 (40.0) 10 (40.0) 3 (12.0) 2 (8.0) 25 (100.0) 16 (64.0)
1 (2.0) 23 (46.0) 16 (32.0) 10 (20.0) 50 (100.0) 21 (42.0)
7 (28.0) 12 (48.0) 6 (24.0)
16 (32.0) 24 (48.0) 10 (20.0)
16 (64.0) 7 (28.0) 2 (8.0) 25 9.9 (N ¼ 21)
0.8199 28 (56.0) 16 (32.0) 6 (12.0) 27 6.6 (N ¼ 47) 0.4401
0.8511 0.7073 0.6902 0.1983 1.0000 0.2752 0.0013
0.0001
e 0.0724 0.8994
treatment-related outcomes. A randomized, controlled trial conducted by Albert and colleagues focused on the impact of antifungal treatment on inflammatory and immune function profiles in patients with VAP and Candida colonization; additional endpoints in this study included clinical outcomes such as mortality and hospital and ICU LOS.13 No statistically significant differences were observed in clinical endpoints between patients who received
Table 2 Primary and secondary outcomes of patients with Candida-positive bronchoscopy cultures and clinical suspicion of pneumonia. Outcomes
Treated (N ¼ 25)
Not treated (N ¼ 50)
P value
In-hospital mortality, n (%) Reason for mortality, n (%) Respiratory cause Infectious cause Other Unable to determine Time to mortality from BAL culture, n (%)a Pneumonia clinical cure, n (%)
6 (24.0)
13 (26.0)
0.8511 0.6357
3 (50.0) 2 (33.3) 1 (16.7) 0 10 5.1 2 (8.7) N ¼ 23 28 26.5 15 11.9 20 20.7 4 8.5
9 (69.2) 3 (23.1) 1 (7.7) 0 5 9.7 9 (23.7) N ¼ 38 23 14.6 14 11.3 14.5 13.3 6 7.8
Length of hospital staya Length of ICU staya Duration of mechanical ventilationa Ventilator-free daysa a
Days, median SD.
Treated (N ¼ 25)
Not treated (N ¼ 50)
P value
20 (80.0) 1 (4.0) 5 (20.0)
41 (82.0) 0 13 (26.0)
1.0000 0.3333 0.5663
1 (4.0) 1 (4.0) 1 (4.0) 0 0 0 2 (8.0) 0 21 (84.0) 15 7.7
2 (4.0) 1 (2.0) 2 (4.0) 1 (2.0) 1 (2.0) 2 (4.0) 3 (6.0) 3 (6.0) 43 (86.0) 10 9.2
1.0000 1.0000 1.0000 1.0000 1.0000 0.5495 1.0000 0.5465 1.0000 0.1541
MSSA ¼ Methicillin-susceptible Staphylococcus aureus; MRSA ¼ Methicillinresistant Staphylococcus aureus.
antifungal therapy or placebo, however this study was considered to be underpowered. Overall, the findings of our retrospective study were consistent with those from Albert and colleagues showing an observed lack of benefit with antifungal treatment.13 Most patients included in the treated group were admitted to our institution in 2011 or 2012, which may indicate changes in prescribing practices in more recent years and could be attributed to the implementation of an antimicrobial stewardship program in 2013. Regardless, prescribing of antifungal agents in this clinical setting continues to be common practice for some physicians based on the results of our institutional survey conducted in 2014 (unpublished results). In our study, patients in the surgical trauma ICU were more likely to receive antifungal agents. These findings were also consistent with our survey results in which trauma and surgical critical care physicians self-reported more antifungal prescribing compared to other physician groups. The subgroup analysis showing similar average APACHE II scores between patients in the medical ICU and surgical trauma ICU suggest severity of illness was similar among the different ICUs. Additionally, mortality between patients residing in these ICUs was found to be similar despite higher prescribing of antifungals in the surgical ICU group. Aside from the differences in ICU location, other patient baseline characteristics (including CPIS and APACHE II scores) between treated and non-treated patients were evenly distributed suggesting our groups were similar in regards to their severity of illness. The median APACHE II scores correlate with in-hospital mortality rates of 40e55%, therefore the observed mortality rates of 24% in the treated group and 26% in the non-treated group were lower than predicted. Of those who experienced in-hospital mortality, there were no differences between groups in terms of cause of death, the majority of which were respiratory or infectious in
Table 4 Antifungal characteristics of treated patients with Candida-positive bronchoscopy cultures and clinical suspicion of pneumonia. Antifungal information
0.2162 0.1822 0.4409 0.2528 0.0755 0.4766
Antifungal agent used, n (%) Fluconazole Caspofungin Fluconazole changed to caspofungin Caspofungin changed to fluconazole Duration of antifungal treatment, days, median SD Time from BAL to start of antifungal treatment, days, median SD
Treated (N ¼ 25) 12 (48.0) 9 (36.0) 3 (12.0) 1 (4.0) 7 4.8 2 1.7
M. Griffin et al. / Heart & Lung xxx (2016) 1e6
nature. Other etiologies for mortality included one paraesophageal hernia and one brain herniation, which were unrelated to the presumed pneumonia in these patients. Despite less than 30% of patients in both groups having bacteria isolated from their BAL, approximately 85% were started on antibiotic therapy. Of the patients that solely had Candida isolated from their BAL, 64% in the treated group and 60% in the non-treated group were started on antibiotics. There are several risk factors associated with the development of invasive fungal infections in immunocompetent, critically ill patients with Candida colonization which may have influenced antifungal prescribing; one of these risk factors is the use of broad-spectrum antibiotics.14 Parenteral nutrition, major surgery, central venous catheters and high APACHE II scores are also risks factors in these patient populations. This study did not account for all of these risk factors; however antibiotic use, baseline APACHE II scores and rates of post-surgical patients were similar between groups. Regardless, we acknowledge that prescribers may have initiated antifungal therapy based on the anticipated risk of general disseminated fungal disease, not only candidal pneumonia. There is a lack of data to suggest an appropriate duration of therapy to treat Candida pneumonia due to the extremely low incidence of the disease. Extrapolating from the IDSA guidelines for treatment of invasive candidiasis in non-neutropenic patients, it is reasonable to discontinue treatment after approximately five days in individuals who do not clinically respond.6 Alternatively, for those who do experience clinical improvement, suggested duration of antifungal therapy is approximately two weeks.6 Patients in the treated group were started on antifungal therapy within two days after the BAL culture and continued to receive treatment for a median duration of seven days. These prescribing patterns suggest that treatment was initiated in response to candidal growth within the BAL but presumably, due to lack of clinical improvement, was discontinued prior to the full treatment period of two weeks. Additionally, preferred empiric treatment for suspected invasive candidiasis in patients who are non-neutropenic is an echinocandin, such as caspofungin, with fluconazole remaining an appropriate alternative in patients without azole-resistant Candida colonization.6 For both regimens, patients should receive a loading dose followed by subsequent lower doses for the remainder of treatment. The majority of patients had C. albicans isolated from their BAL which is generally sensitive to fluconazole and caspofungin, although susceptibilities of these isolates were not reported. In general, caspofungin was initiated appropriately with a loading dose of 70 mg followed by subsequent 50 mg dosing. Fluconazole dosing was less consistent with the majority of patients initiated on treatment without a load, and doses varied between 200 mg and 400 mg daily. Rates of pneumonia clinical cure, based on normal WBC count and temperature at day 8 from BAL, were low in both groups. The observed short duration of antifungal treatment and differences in dosing strategies may have influenced clinical cure rates; however, these lower rates were not unexpected since factors other than fungal infection can impact these nonspecific markers in critically ill patients. These findings are consistent with those published by Raman and colleagues who concluded discontinuation of antimicrobial therapy did not impact clinical outcomes despite persistence of pneumonia clinical signs and symptoms such as elevated WBC and temperature.15 Therefore, inappropriate treatment of Candida pneumonia was not likely a reason for respiratory or infection-related mortality observed in the treatment group. Finally, Candida colonization of the respiratory tract has been associated with higher rates of development of multi-drug resistant organisms and, in one study, worse outcomes in patients with VAP, including prolonged hospital and ICU LOS.16e18 Pseudomonas
5
aeruginosa, which was not present on the initial BAL, was isolated in a total of three patients; two of these isolates were in the treated group and one in the non-treated group. Of these three patients, one in the treated group experienced in-hospital mortality. Due to the low number of patients, we were unable to draw conclusions regarding the treatment of Candida resulting in fewer patients developing pseudomonal infections. In addition, there was no incidence of fluconazole-resistant Candida or non-albicans Candida on repeat culture. However, only 34.7% of patients had a repeat respiratory culture in the 28 days following the original BAL. We recognize there are several limitations to our study. First, this was a retrospective study with a small sample size of immunocompetent patients, which could limit the generalizability. Also, the study was underpowered and we may have been unable to detect statistically significant effects. Second, there may have been confounding factors that were not accounted for that impacted patient outcomes. Additionally, there were limitations with data collection because CPIS and APACHE II scores could not be calculated for all patients due to missing data within electronic medical records. Lastly, we recognize the incidence of true Candida pneumonia is quite low and therefore we cannot conclude the patients included in this study had Candida pneumonia. However, since true diagnosis via lung tissue biopsy is rarely done in clinical practice, the results of this study based on presumed pneumonia are still valuable. Conclusion This is the first study that focuses on treatment-related outcomes for presumed Candida pneumonia. We did not observe a difference in mortality or clinical outcomes in critically-ill, immunocompetent patients treated with antifungal agents, although the study was underpowered. However, in alignment with the IDSA guidelines, our results do not support starting antifungal therapy based on a positive BAL alone in critically ill, immunocompetent patients. Acknowledgements Angela Kashyap, PharmD (data collection assistance). References 1. Wisplinghoff H, Bischoff T, Tallent SM, et al. Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study. Clin Infect Dis. 2004;39:309e317. 2. Baum GL. The significance of Candida species in human sputum. N Eng J Med. 1960;263:70e73. 3. Meersseman W, Lagrou K, Spriet I, et al. Significance of the isolation of Candida species from airway samples in critically ill patients: a prospective, autopsy study. Intensive Care Med. 2009;35:1526e1531. 4. El-Ebiary M, Torres A, Fabregas N, et al. Significance of isolation of Candida species from respiratory samples in critically ill, non-neutropenic patients. Am J Respir Crit Care Med. 1997;156:583e590. 5. Rello J, Esandi ME, Diaz E, et al. The role of Candida sp isolated from bronchoscopic samples in nonneutropenic patients. Chest. 1998;114:146e149. 6. Pappas PG, Kauffman CA, Andes DR, et al. Clinical practice guidelines for the management of candidiasis: 2016 update by the Infectious Diseases Society of America. Clin Infect Dis. 2016;62(4):e1ee50. 7. Azoulay E, Cohen Y, Zahar JR, et al. Practices in non-neutropenic ICU patients with Candida-positive airway specimens. Intensive Care Med. 2004;30: 1384e1389. 8. Anada-Rajah MR, Slavin MA, Thursky KT. The case for antifungal stewardship. Curr Opin Infect Dis. 2012;25:107e115. 9. Lortholary O, Desnos-Ollivier M, Sitbon K, et al. Recent exposure to caspofungin or fluconazole influences the epidemiology of candidemia: a prospective multicenter study involving 2441 patients. Antimicrob Agents Chemother. 2011; 55(2):532e538. 10. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap) e a metadata-driven methodology and workflow process
6
11. 12.
13.
14.
M. Griffin et al. / Heart & Lung xxx (2016) 1e6 for providing translational research informatics support. J Biomed Inf. 2009 Apr;42(2):377e381. Schnabel RM, Linssen CF, Guion N, et al. Candida pneumonia in intensive care unit? Open Forum Infect Dis. 2014;1(1):1e6. Delisle M-S, Williamson DR, Perreault MM, et al. The clinical significance of Candida colonization of respiratory tract secretions in critically ill patients. J Crit Care. 2008;23:11e17. Albert M, Williamson D, Muscedere J, et al. Candida in the respiratory tract secretions of critically ill patients and the impact of antifungal treatment: a randomized placebo controlled pilot trial (CANTREAT study). Intensive Care Med. 2014;40:1313e1322. Muskett H, Shahin J, Eyres G, et al. Risk factors for invasive fungal disease in critically ill adult patients; a systematic review. Crit Care. 2011;15:R287.
15. Raman K, Nailor MD, Nicolau DP, et al. Early antibiotic discontinuation in patients with clinically suspected ventilator-associated pneumonia and negative quantitative bronchoscopy cultures. Crit Care Med. 2013;41:1656e1663. 16. Hamet M, Pavon A, Dalle F, et al. Candida spp. airway colonization could promote antibiotic-resistant bacteria selection in patients with suspected ventilator-associated pneumonia. Intensive Care Med. 2012;38: 1272e1279. 17. Azoulay E, Timsit J, Tafflet M, et al. Candida colonization of the respiratory tract and subsequent Pseudomonas ventilator-associated pneumonia. Chest. 2006;129:110e117. 18. Nseir S, Jozefowicz E, Cavestri B, et al. Impact of antifungal treatment on CandidaePseudomonas interaction: a preliminary retrospective caseecontrol study. Intensive Care Med. 2007;33(1):137e142.
Contents lists available at ScienceDirect
Heart & Lung journal homepage: www.heartandlung.org
Antifungal use in immunocompetent, critically ill patients with pneumonia does not improve clinical outcomes Meagan Griffin, PharmD a, Desiree E. Kosmisky, PharmD c, Megan A. Templin, MPH, MS c, Toan Huynh, MD b, Lewis H. McCurdy III, MD b, Timothy R. Pasquale, PharmD b, Kelly E. Martin, PharmD b, * a
Memorial Sloan Kettering Cancer Center, 1275 York Avenue, New York, NY 10065, USA Carolinas Medical Center, 1000 Blythe Boulevard, Charlotte, NC 28203, USA c Carolinas HealthCare System, Charlotte, NC 28203, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 18 April 2016 Received in revised form 1 August 2016 Accepted 2 August 2016 Available online xxx
Purpose: To determine if treating bronchoalveolar lavage (BAL) culture-positive patients with antifungal therapy impacted mortality compared to not treating due to presumed colonization. Methods: We conducted a retrospective study of immunocompetent, critically ill adult patients from 2010 to 2014. Patients with a BAL culture-positive for Candida or unspeciated yeast and a clinical suspicion of pneumonia were included. The treatment group received an antifungal agent for at least 5 days, and the control group received either no antifungal therapy or an antifungal agent for less than 48 h. Recruitment occurred in a 2:1 ratio of untreated versus treated patients. Results: Seventy-five patients were included. In-hospital mortality was similar between treated and untreated groups (24% vs. 26%, P ¼ 0.85). Length of stay and duration of mechanical ventilation also did not differ between the two groups. Conclusion: We did not observe a difference in mortality or clinical outcomes in patients treated with antifungal agents. Presumptive antifungal therapy for BAL-positive Candida or yeast in immunocompetent patients did not result in improved clinical outcomes. Ó 2016 Elsevier Inc. All rights reserved.
Keywords: Candida Yeast Bronchoscopy Antifungal agents Pneumonia
Introduction Candida species account for 9% of hospital-acquired bloodstream infections and candidemia is associated with a crude mortality of up to 47%.1 However, Candida albicans, the most common species of yeast, is an oropharyngeal colonizer in up to 20% of healthy individuals and 55% of hospitalized patients.2 The diagnosis of candidal pneumonia in non-neutropenic patients is extremely rare. Indeed, candidal pneumonia requires histopathologic confirmation with findings of yeast cells or pseudohyphae on lung biopsy, which is a practice not routinely performed.3 Candida isolated from
Conflict of interest: The authors whose names are listed immediately below certify that they have no disclosures in the subject matter or materials discussed in this manuscript. Meagan Griffin, Desiree E. Kosmisky, Megan A. Templin, Toan Huynh, Kelly E. Martin. The authors whose names are listed immediately below are part of the Cubist Speaker’s Bureau. Timothy R. Pasquale, Lewis H. McCurdy III. * Corresponding author. E-mail address: [email protected] (K.E. Martin). 0147-9563/$ e see front matter Ó 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.hrtlng.2016.08.002
bronchoscopic specimens is usually thought to be non-pathogenic.2 Several studies have confirmed that isolation of Candida from a bronchoalveolar lavage (BAL), although common in hospitalized patients, is rarely an active source of pneumonia.3e5 As such, current guidelines for candidiasis do not recommend the initiation of antifungal treatment based on a positive BAL alone due to the frequent rate of colonization, the poor predictive value of this finding, and the rarity of candidal pneumonia.6 Despite these guideline recommendations, providers frequently prescribe antifungal agents for critically ill, febrile patients with no clear evidence of invasive fungal infection. Indeed, our recent unpublished institutional survey of 64 providers from trauma surgical critical care, pulmonary critical care, and infectious diseases demonstrated significant variability in self-reported prescribing practices of antifungal therapy in critically ill patients. Up to 24% of respondents reported the empiric use of antifungal therapy for clinical suspicion of pneumonia, and 42% would initiate antifungal treatment of BAL cultures positive for C. albicans. Overall, trauma surgical critical care providers were most likely to consider Candida to be a cause of pneumonia and to prescribe antifungals. By contrast, infectious diseases physicians were the least likely to
2
M. Griffin et al. / Heart & Lung xxx (2016) 1e6
prescribe antifungal therapy for pneumonia. The practice of prescribing antifungals in this clinical setting is not limited to our institution. Our survey results are consistent with a study assessing antifungal prescribing practices in critically ill, non-neutropenic patients in France, where 24% of intensivists reported prescribing antifungal therapy for patients with chronic obstructive pulmonary disease requiring mechanical ventilation with a tracheal aspirate positive for Candida species.7 The overuse of antifungal therapy can lead to unintended consequences including adverse effects and increased resistance.8 Previous azole exposure is a risk factor for subsequent Candida infections with decreased susceptibility profiles, such as Candida glabrata and Candida krusei.9 The Infectious Diseases Society of America (IDSA) guidelines recommend the empiric use of an echinocandin in patients with recent treatment with an azole for this very reason.6 The purpose of the study was to determine if treating patients with antifungal therapy for a BAL culture positive for Candida or unspeciated yeast had an impact on mortality compared to not treating due to presumed colonization. The primary outcome was a comparison of in-hospital mortality between the treatment and non-treatment groups. Secondary outcomes included hospital length of stay (LOS), intensive care unit (ICU) LOS, duration of mechanical ventilation, and time to mortality from BAL culture. Methods Design This was a single-center, retrospective chart review of critically ill adult patients admitted to one of three ICUs (medical ICU, surgical trauma ICU and neurosurgical ICU) at an academic medical center from January 1, 2010 to July 31, 2014. This study received Institutional Review Board approval. Inclusion criteria consisted of age 18 years old, mechanical ventilation at the time of BAL and at least one BAL positive for Candida or unspeciated yeast during ICU admission. Patients who met criteria for clinical suspicion of nosocomial pneumonia on the day of BAL were included. These criteria consisted of the presence of infiltrate on chest radiograph or chest computerized tomography (CT) based on radiology reports, new onset purulent sputum, changes in secretions or secretions noted on BAL, and either a temperature of 38 C or 36 C and/or a white blood cell (WBC) of 12,000 or 4000 cells/mm3. Patients were excluded if they were severely immunocompromised, defined as patients with a history of hematopoietic-stem cell transplantation, chemotherapy in the last 30 days, solid organ transplantation, human immunodeficiency virus (HIV) with a CD4 T-lymphocyte count of <200 cells/mm3, or daily corticosteroid therapy with a dose equivalent to 20 mg of prednisone for 14 days. Patients who received antifungals for reasons other than pneumonia were excluded using the following criteria: medical record documentation of antifungal use for other indications, presence of positive blood, fluid, or urine cultures with Candida species 7 days prior to or during antifungal therapy, receiving an antifungal agent other than fluconazole or an echinocandin, and receiving antifungal treatment duration of >48 hours but <5 days. Patients were also excluded if they were diagnosed with community-acquired pneumonia (CAP), pregnant, had a BAL growing a carbapenem-resistant or pan-resistant organism, previously included in the study, or if there was an absence of data to assess clinical suspicion of pneumonia. Patients with bacteria isolated in a BAL culture could be included if the organism was adequately treated with an antibiotic the organism was susceptible to for at least seven days.
Variables and outcome measures Data included age, gender, race, weight, and comorbidities. Characteristics related to pneumonia consisted of presumed pneumonia, radiographic evidence of pneumonia by either chest radiograph or CT scan, and ratio of partial pressure of arterial oxygen and fraction of inspired oxygen (PaO2/FiO2). Disease severity scores were assessed by Acute Physiology and Chronic Health Evaluation (APACHE) II on day of ICU admission and Clinical Pulmonary Infection Score (CPIS) on day of BAL. Microbiologic data included bacterial species isolated from the respiratory tract and bronchoscopy findings, including purulent sputum. Cultures within 28 days were also reviewed for presence of fluconazole-resistant, or non-albicans Candida species and Pseudomonas species. WBC count and maximum temperature were collected on the day of BAL and persistence of these factors were assessed on day 8 from BAL. Maximum heart rate, respiratory rate or partial pressure of arterial carbon dioxide (PaCO2), systolic blood pressure, and lactate were collected on the day of BAL. Pertinent antimicrobial therapy included duration of concurrent antibiotics, duration of antifungal therapy, time from bronchoscopy to initiation of antifungal therapy, antifungal agent prescribed (fluconazole or echinocandin), fluconazole dose and route, and echinocandin doses. Clinical endpoints consisted of in-hospital mortality, time to mortality from BAL culture, hospital length of stay (LOS), ICU LOS, duration of mechanical ventilation, and cause of death per death records documented within patient charts. Study data were collected and managed using REDCap (Research Electronic Data Capture) tools hosted at Carolinas HealthCare System.10 REDCap is a secure, web-based application which provides an interface for validated data entry, audit trails for tracking data manipulation and export procedures, and automated export procedures for data downloads. Statistical analysis Data were analyzed using descriptive statistics consisting of means, medians, and percentages. The primary analysis was a chisquare test comparing in-hospital mortality for the two treatment groups and the corresponding 95% confidence intervals on the difference in the proportions. Secondary outcomes were analyzed with the non-parametric Wilcoxon rank-sum test. SAS Enterprise GuideÒ, version 5.1. A two-tailed P-value of less than 0.05 was considered statistically significant. Results Due to low enrollment in the treatment group, the IRB-approved protocol was modified to include patients in a 2 to 1 ratio in reverse chronological order until a total of 75 patients were included. There were a total of 25 patients in the treated group and 50 patients in the non-treated group (Fig. 1). A total of 224 patients were excluded from the study, of which 22.8% were found to have evidence of Candida or unspeciated yeast in other cultures. The majority of patients were Caucasian males, with a median age of 59 years old (Table 1). Most patients had clinical suspicion of ventilatorassociated pneumonia (VAP) with a median APACHE II score of 25 in the treated group and 27 in the non-treated group. Of note, APACHE II scores were not able to be calculated for four patients in the treated group and three patients in the non-treated group due to lack of data. Differences between groups occurred in relation to ICU-specialty, with significantly more treated patients residing in the surgical trauma ICU compared to the other ICUs (P ¼ 0.0013). A subgroup analysis of all patients in the medical ICU and surgical trauma ICU showed average APACHE II scores of 34 and 29 respectively (P ¼ 0.2684). Additionally, mortality between medical
M. Griffin et al. / Heart & Lung xxx (2016) 1e6
3
316 posiƟve paƟent bronchoscopies screened
224 Excluded: 62 Previously included paƟents 51 Candida/yeast other cultures 30 No infiltrate on CXR/CT 25 Immunocompromised 13 Diagnosed CAP 11 AnƟfungal therapy > 48 hours but < 5 days 10 Didn't meet clinical criteria 8 Not mechanically venƟlated
17 non-treated paƟents excluded when goal inclusion of 50 non-treated paƟents met
5 Already on anƟfungal agent 4 Bacterial pathogens not treated properly 3 Absence of data to assess pneumonia 1 Age < 18 years old 1 Use of other anƟfungal agent
Treated group:
Non-treated group:
25 paƟents
50 paƟents
Fig. 1. Patient enrollment.
ICU and surgical trauma ICU patients was found to be similar (37.5% vs 17.14%, P ¼ 0.0604). Overall, no statistically significant difference was observed between the treated and non-treated groups for the primary endpoint of in-hospital mortality (24% vs. 26%, P ¼ 0.8511), or the secondary endpoints of time to mortality from BAL culture, pneumonia clinical cure at day 8 from BAL, hospital LOS, ICU LOS, and duration of mechanical ventilation (Table 2). Additionally, the median number of ventilator-free days was 4 in the treated group and 6 in the nontreated group (P ¼ 0.4766). C. albicans accounted for the majority of Candida isolates in both groups (Table 3). Additionally, unspeciated yeast was isolated in 20% of treated and 32% of non-treated patients. There were a total of 26 patients who had a repeat BAL within 28 days; Pseudomonas was isolated in three patients. There was no incidence of fluconazole-resistant Candida or non-albicans Candida on any repeat respiratory culture. Forty-eight percent of patients received fluconazole for treatment of their presumed Candida pneumonia, 36% received caspofungin and 16% switched between both agents (Table 4). Of the patients who received fluconazole, 93.8% received intravenous therapy. Seventy-five percent received a dose of 400 mg daily and 25% received a dose of 200 mg daily. A total of 18.8% of patients received a fluconazole loading dose that was twice the maintenance dose. Of the patients who received caspofungin,
76.9% were treated with a 70 mg loading dose followed by 50 mg daily, 15.4% received 50 mg daily without a loading dose and 7.7% of patients received a 70 mg loading dose followed by 35 mg daily to adjust for hepatic insufficiency. Discussion To our knowledge, this is the first study conducted to focus on treatment-related outcomes of patients with Candida-positive BALs. Our results demonstrated that treatment of Candida-positive BALs with antifungal therapy had no significant impact on inhospital mortality, time to mortality, hospital LOS, or duration of mechanical ventilation compared to patients who were not treated. Several previous studies have investigated the impact of Candida colonization in critically ill patients. Schnabel and colleagues examined whether Candida pneumonia was a clinical concern within the critically ill population.11 Of 701 BAL samples analyzed, only 0.7% correlated with a proven candidal pneumonia. These findings suggested that true fungal pneumonia in this setting is rare despite frequent colonization.11 In a retrospective analysis of patients with diagnosed VAP conducted by Delisle and colleagues, colonization of Candida within the respiratory tract was associated with increased hospital mortality and increased hospital LOS.12 However, neither of these studies were focused on
4
M. Griffin et al. / Heart & Lung xxx (2016) 1e6
Table 1 Baseline characteristics of patients with Candida-positive bronchoscopy cultures and clinical suspicion of pneumonia.
Table 3 Microbial information of patients with Candida-positive bronchoscopy cultures and clinical suspicion of pneumonia.
Characteristics
Treated (N ¼ 25)
Not treated (N ¼ 50)
P value
Microbial information
Age, years, median SD Weight, kg, median SD Gender, male, n (%) Race American Indian or Alaskan Native Black or African American Native Hawaiian or Pacific Islander White Unknown Comorbid conditions, n (%) Diabetes COPD Renal failure Malignancy Liver disease Post-surgical Intensive care unit (ICU), n (%) Medical ICU Surgical trauma ICU Neurosurgical ICU Year of admission, n (%) 2011 2012 2013 2014 SIRS criteria met Severe sepsis criteria met CPIS suggestive of pneumonia, n (%) Yes No Unable to determine Type of pneumonia, n (%) Ventilator-associated Hospital-acquired Healthcare-associated APACHE II score, median SD
60 16.5 90 21.7 18 (72.0)
59 15.5 82 27.8 30 (60.0)
0.9140 0.9418 0.3074 0.7521
1 (4.0)
0
3 (12.0) 0
7 (14.0) 1 (2.0)
20 (80.0) 1 (4.0)
39 (78.0) 3 (6.0)
6 7 6 6 3 9
(24.0) (28.0) (24.0) (24.0) (12.0) (36.0)
13 (26.0) 12 (24.0) 10 (20.0) 6 (12.0) 7 (14.0) 12 (24.0)
Type of Candida isolated, n (%) Candida albicans Candida tropicalis Bacteria isolated from BAL, n (%) Type of bacteria isolated, n (%) Pseudomonas aeruginosa Klebsiella spp. Haemophilus influenzae Acinetobacter spp. Streptococcus pneumoniae MSSA MRSA Other Antibiotics started, n (%) Duration of antibiotics, days, median SD
5 (20.0) 19 (76.0) 1 (4.0)
27 (54.0) 16 (32.0) 7 (14.0)
10 (40.0) 10 (40.0) 3 (12.0) 2 (8.0) 25 (100.0) 16 (64.0)
1 (2.0) 23 (46.0) 16 (32.0) 10 (20.0) 50 (100.0) 21 (42.0)
7 (28.0) 12 (48.0) 6 (24.0)
16 (32.0) 24 (48.0) 10 (20.0)
16 (64.0) 7 (28.0) 2 (8.0) 25 9.9 (N ¼ 21)
0.8199 28 (56.0) 16 (32.0) 6 (12.0) 27 6.6 (N ¼ 47) 0.4401
0.8511 0.7073 0.6902 0.1983 1.0000 0.2752 0.0013
0.0001
e 0.0724 0.8994
treatment-related outcomes. A randomized, controlled trial conducted by Albert and colleagues focused on the impact of antifungal treatment on inflammatory and immune function profiles in patients with VAP and Candida colonization; additional endpoints in this study included clinical outcomes such as mortality and hospital and ICU LOS.13 No statistically significant differences were observed in clinical endpoints between patients who received
Table 2 Primary and secondary outcomes of patients with Candida-positive bronchoscopy cultures and clinical suspicion of pneumonia. Outcomes
Treated (N ¼ 25)
Not treated (N ¼ 50)
P value
In-hospital mortality, n (%) Reason for mortality, n (%) Respiratory cause Infectious cause Other Unable to determine Time to mortality from BAL culture, n (%)a Pneumonia clinical cure, n (%)
6 (24.0)
13 (26.0)
0.8511 0.6357
3 (50.0) 2 (33.3) 1 (16.7) 0 10 5.1 2 (8.7) N ¼ 23 28 26.5 15 11.9 20 20.7 4 8.5
9 (69.2) 3 (23.1) 1 (7.7) 0 5 9.7 9 (23.7) N ¼ 38 23 14.6 14 11.3 14.5 13.3 6 7.8
Length of hospital staya Length of ICU staya Duration of mechanical ventilationa Ventilator-free daysa a
Days, median SD.
Treated (N ¼ 25)
Not treated (N ¼ 50)
P value
20 (80.0) 1 (4.0) 5 (20.0)
41 (82.0) 0 13 (26.0)
1.0000 0.3333 0.5663
1 (4.0) 1 (4.0) 1 (4.0) 0 0 0 2 (8.0) 0 21 (84.0) 15 7.7
2 (4.0) 1 (2.0) 2 (4.0) 1 (2.0) 1 (2.0) 2 (4.0) 3 (6.0) 3 (6.0) 43 (86.0) 10 9.2
1.0000 1.0000 1.0000 1.0000 1.0000 0.5495 1.0000 0.5465 1.0000 0.1541
MSSA ¼ Methicillin-susceptible Staphylococcus aureus; MRSA ¼ Methicillinresistant Staphylococcus aureus.
antifungal therapy or placebo, however this study was considered to be underpowered. Overall, the findings of our retrospective study were consistent with those from Albert and colleagues showing an observed lack of benefit with antifungal treatment.13 Most patients included in the treated group were admitted to our institution in 2011 or 2012, which may indicate changes in prescribing practices in more recent years and could be attributed to the implementation of an antimicrobial stewardship program in 2013. Regardless, prescribing of antifungal agents in this clinical setting continues to be common practice for some physicians based on the results of our institutional survey conducted in 2014 (unpublished results). In our study, patients in the surgical trauma ICU were more likely to receive antifungal agents. These findings were also consistent with our survey results in which trauma and surgical critical care physicians self-reported more antifungal prescribing compared to other physician groups. The subgroup analysis showing similar average APACHE II scores between patients in the medical ICU and surgical trauma ICU suggest severity of illness was similar among the different ICUs. Additionally, mortality between patients residing in these ICUs was found to be similar despite higher prescribing of antifungals in the surgical ICU group. Aside from the differences in ICU location, other patient baseline characteristics (including CPIS and APACHE II scores) between treated and non-treated patients were evenly distributed suggesting our groups were similar in regards to their severity of illness. The median APACHE II scores correlate with in-hospital mortality rates of 40e55%, therefore the observed mortality rates of 24% in the treated group and 26% in the non-treated group were lower than predicted. Of those who experienced in-hospital mortality, there were no differences between groups in terms of cause of death, the majority of which were respiratory or infectious in
Table 4 Antifungal characteristics of treated patients with Candida-positive bronchoscopy cultures and clinical suspicion of pneumonia. Antifungal information
0.2162 0.1822 0.4409 0.2528 0.0755 0.4766
Antifungal agent used, n (%) Fluconazole Caspofungin Fluconazole changed to caspofungin Caspofungin changed to fluconazole Duration of antifungal treatment, days, median SD Time from BAL to start of antifungal treatment, days, median SD
Treated (N ¼ 25) 12 (48.0) 9 (36.0) 3 (12.0) 1 (4.0) 7 4.8 2 1.7
M. Griffin et al. / Heart & Lung xxx (2016) 1e6
nature. Other etiologies for mortality included one paraesophageal hernia and one brain herniation, which were unrelated to the presumed pneumonia in these patients. Despite less than 30% of patients in both groups having bacteria isolated from their BAL, approximately 85% were started on antibiotic therapy. Of the patients that solely had Candida isolated from their BAL, 64% in the treated group and 60% in the non-treated group were started on antibiotics. There are several risk factors associated with the development of invasive fungal infections in immunocompetent, critically ill patients with Candida colonization which may have influenced antifungal prescribing; one of these risk factors is the use of broad-spectrum antibiotics.14 Parenteral nutrition, major surgery, central venous catheters and high APACHE II scores are also risks factors in these patient populations. This study did not account for all of these risk factors; however antibiotic use, baseline APACHE II scores and rates of post-surgical patients were similar between groups. Regardless, we acknowledge that prescribers may have initiated antifungal therapy based on the anticipated risk of general disseminated fungal disease, not only candidal pneumonia. There is a lack of data to suggest an appropriate duration of therapy to treat Candida pneumonia due to the extremely low incidence of the disease. Extrapolating from the IDSA guidelines for treatment of invasive candidiasis in non-neutropenic patients, it is reasonable to discontinue treatment after approximately five days in individuals who do not clinically respond.6 Alternatively, for those who do experience clinical improvement, suggested duration of antifungal therapy is approximately two weeks.6 Patients in the treated group were started on antifungal therapy within two days after the BAL culture and continued to receive treatment for a median duration of seven days. These prescribing patterns suggest that treatment was initiated in response to candidal growth within the BAL but presumably, due to lack of clinical improvement, was discontinued prior to the full treatment period of two weeks. Additionally, preferred empiric treatment for suspected invasive candidiasis in patients who are non-neutropenic is an echinocandin, such as caspofungin, with fluconazole remaining an appropriate alternative in patients without azole-resistant Candida colonization.6 For both regimens, patients should receive a loading dose followed by subsequent lower doses for the remainder of treatment. The majority of patients had C. albicans isolated from their BAL which is generally sensitive to fluconazole and caspofungin, although susceptibilities of these isolates were not reported. In general, caspofungin was initiated appropriately with a loading dose of 70 mg followed by subsequent 50 mg dosing. Fluconazole dosing was less consistent with the majority of patients initiated on treatment without a load, and doses varied between 200 mg and 400 mg daily. Rates of pneumonia clinical cure, based on normal WBC count and temperature at day 8 from BAL, were low in both groups. The observed short duration of antifungal treatment and differences in dosing strategies may have influenced clinical cure rates; however, these lower rates were not unexpected since factors other than fungal infection can impact these nonspecific markers in critically ill patients. These findings are consistent with those published by Raman and colleagues who concluded discontinuation of antimicrobial therapy did not impact clinical outcomes despite persistence of pneumonia clinical signs and symptoms such as elevated WBC and temperature.15 Therefore, inappropriate treatment of Candida pneumonia was not likely a reason for respiratory or infection-related mortality observed in the treatment group. Finally, Candida colonization of the respiratory tract has been associated with higher rates of development of multi-drug resistant organisms and, in one study, worse outcomes in patients with VAP, including prolonged hospital and ICU LOS.16e18 Pseudomonas
5
aeruginosa, which was not present on the initial BAL, was isolated in a total of three patients; two of these isolates were in the treated group and one in the non-treated group. Of these three patients, one in the treated group experienced in-hospital mortality. Due to the low number of patients, we were unable to draw conclusions regarding the treatment of Candida resulting in fewer patients developing pseudomonal infections. In addition, there was no incidence of fluconazole-resistant Candida or non-albicans Candida on repeat culture. However, only 34.7% of patients had a repeat respiratory culture in the 28 days following the original BAL. We recognize there are several limitations to our study. First, this was a retrospective study with a small sample size of immunocompetent patients, which could limit the generalizability. Also, the study was underpowered and we may have been unable to detect statistically significant effects. Second, there may have been confounding factors that were not accounted for that impacted patient outcomes. Additionally, there were limitations with data collection because CPIS and APACHE II scores could not be calculated for all patients due to missing data within electronic medical records. Lastly, we recognize the incidence of true Candida pneumonia is quite low and therefore we cannot conclude the patients included in this study had Candida pneumonia. However, since true diagnosis via lung tissue biopsy is rarely done in clinical practice, the results of this study based on presumed pneumonia are still valuable. Conclusion This is the first study that focuses on treatment-related outcomes for presumed Candida pneumonia. We did not observe a difference in mortality or clinical outcomes in critically-ill, immunocompetent patients treated with antifungal agents, although the study was underpowered. However, in alignment with the IDSA guidelines, our results do not support starting antifungal therapy based on a positive BAL alone in critically ill, immunocompetent patients. Acknowledgements Angela Kashyap, PharmD (data collection assistance). References 1. Wisplinghoff H, Bischoff T, Tallent SM, et al. Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study. Clin Infect Dis. 2004;39:309e317. 2. Baum GL. The significance of Candida species in human sputum. N Eng J Med. 1960;263:70e73. 3. Meersseman W, Lagrou K, Spriet I, et al. Significance of the isolation of Candida species from airway samples in critically ill patients: a prospective, autopsy study. Intensive Care Med. 2009;35:1526e1531. 4. El-Ebiary M, Torres A, Fabregas N, et al. Significance of isolation of Candida species from respiratory samples in critically ill, non-neutropenic patients. Am J Respir Crit Care Med. 1997;156:583e590. 5. Rello J, Esandi ME, Diaz E, et al. The role of Candida sp isolated from bronchoscopic samples in nonneutropenic patients. Chest. 1998;114:146e149. 6. Pappas PG, Kauffman CA, Andes DR, et al. Clinical practice guidelines for the management of candidiasis: 2016 update by the Infectious Diseases Society of America. Clin Infect Dis. 2016;62(4):e1ee50. 7. Azoulay E, Cohen Y, Zahar JR, et al. Practices in non-neutropenic ICU patients with Candida-positive airway specimens. Intensive Care Med. 2004;30: 1384e1389. 8. Anada-Rajah MR, Slavin MA, Thursky KT. The case for antifungal stewardship. Curr Opin Infect Dis. 2012;25:107e115. 9. Lortholary O, Desnos-Ollivier M, Sitbon K, et al. Recent exposure to caspofungin or fluconazole influences the epidemiology of candidemia: a prospective multicenter study involving 2441 patients. Antimicrob Agents Chemother. 2011; 55(2):532e538. 10. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap) e a metadata-driven methodology and workflow process
6
11. 12.
13.
14.
M. Griffin et al. / Heart & Lung xxx (2016) 1e6 for providing translational research informatics support. J Biomed Inf. 2009 Apr;42(2):377e381. Schnabel RM, Linssen CF, Guion N, et al. Candida pneumonia in intensive care unit? Open Forum Infect Dis. 2014;1(1):1e6. Delisle M-S, Williamson DR, Perreault MM, et al. The clinical significance of Candida colonization of respiratory tract secretions in critically ill patients. J Crit Care. 2008;23:11e17. Albert M, Williamson D, Muscedere J, et al. Candida in the respiratory tract secretions of critically ill patients and the impact of antifungal treatment: a randomized placebo controlled pilot trial (CANTREAT study). Intensive Care Med. 2014;40:1313e1322. Muskett H, Shahin J, Eyres G, et al. Risk factors for invasive fungal disease in critically ill adult patients; a systematic review. Crit Care. 2011;15:R287.
15. Raman K, Nailor MD, Nicolau DP, et al. Early antibiotic discontinuation in patients with clinically suspected ventilator-associated pneumonia and negative quantitative bronchoscopy cultures. Crit Care Med. 2013;41:1656e1663. 16. Hamet M, Pavon A, Dalle F, et al. Candida spp. airway colonization could promote antibiotic-resistant bacteria selection in patients with suspected ventilator-associated pneumonia. Intensive Care Med. 2012;38: 1272e1279. 17. Azoulay E, Timsit J, Tafflet M, et al. Candida colonization of the respiratory tract and subsequent Pseudomonas ventilator-associated pneumonia. Chest. 2006;129:110e117. 18. Nseir S, Jozefowicz E, Cavestri B, et al. Impact of antifungal treatment on CandidaePseudomonas interaction: a preliminary retrospective caseecontrol study. Intensive Care Med. 2007;33(1):137e142.