Secondarily documented invasive candidiasis is unpredictable using traditional risk factors in non transplant – non-neutropenic adult ICU patients

Abstract

•Intra-abdominal surgery is the sole risk factor differentiating secondary proven invasive candidiasis from other cause of sepsis when empirical systemic antifungal therapy is started in ICU non-neutropenic patients.•Diagnostic tests are urgently needed. Invasive candidiasis (IC) in intensive care units (ICUs) is the most common cause of fungal infection, and it is associated with a high attributable mortality.1Leroy O. Bailly S. Gangneux J.P. Mira J.P. Devos P. Dupont H. et al.Systemic antifungal therapy for proven or suspected invasive candidiasis: the AmarCAND 2 study.Ann Intensive Care. 2016; 6: 2Crossref PubMed Scopus (65) Google Scholar Many risk factors are associated with IC and were integrated in different clinical scores (Candida score, colonization index, Dupont score) but their variable predictive values, far better for the negative predictive value than for the positive one, do not allow use as decision tools for treatment administration.2Leon C. Ostrosky-Zeichner L. Schuster M. What's new in the clinical and diagnostic management of invasive candidiasis in critically ill patients.Intensive Care Med. 2014; 40: 808-819Crossref PubMed Scopus (85) Google Scholar Moreover, the absence of effective diagnostic tools2Leon C. Ostrosky-Zeichner L. Schuster M. What's new in the clinical and diagnostic management of invasive candidiasis in critically ill patients.Intensive Care Med. 2014; 40: 808-819Crossref PubMed Scopus (85) Google Scholar spurs us to recommend and largely prescribe empirical systemic antifungal therapies (SAT) to the ICU patients with risk factors.3Azoulay E. Dupont H. Tabah A. Lortholary O. Stahl J.P. Francais A. et al.Systemic antifungal therapy in critically ill patients without invasive fungal infection*.Crit Care Med. 2012; 40: 813-822Crossref PubMed Scopus (102) Google Scholar, 4Pappas P.G. Kauffman C.A. Andes D.R. Clancy C.J. Marr K.A. Ostrosky-Zeichner L. et al.Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the Infectious Diseases Society of America.Clin Infect Dis. 2015; Google Scholar, 5Eggimann P. Pittet D. Candida colonization index and subsequent infection in critically ill surgical patients: 20 years later.Intensive Care Med. 2014; 40: 1429-1448Crossref PubMed Scopus (96) Google Scholar Empirical SAT is associated with increase in costs6Munoz P. Valerio M. Vena A. Bouza E. Antifungal stewardship in daily practice and health economic implications.Mycoses. 2015; 58: 14-25Crossref PubMed Scopus (72) Google Scholar and a selection of resistant strains,7Alexander B.D. Johnson M.D. Pfeiffer C.D. Jimenez-Ortigosa C. Catania J. Booker R. et al.Increasing echinocandin resistance in Candida glabrata: clinical failure correlates with presence of FKS mutations and elevated minimum inhibitory concentrations.Clin Infect Dis. 2013; 56: 1724-1732Crossref PubMed Scopus (543) Google Scholar and it contributes to the increase in antifungal consumption responsible for modification in fungal species distribution and susceptibility.8Bailly S. Maubon D. Fournier P. Pelloux H. Schwebel C. Chapuis C. et al.Impact of antifungal prescription on relative distribution and susceptibility of Candida spp. - Trends over 10 years.J Infect. 2016; 72: 103-111Abstract Full Text Full Text PDF PubMed Scopus (57) Google Scholar However, secondarily proven invasive candidiasis (SPIC) is only observed in a fraction of those patients who receive an empirical SAT. The objective of this study was to define the risk factors predictive for SPIC onset during empirical SAT, by accounting for center and patient variables in a large prospective cohort. The subpopulation of non-transplant non-neutropenic adult ICU patients with an empirical SAT during their ICU stay (≥ 48 h post-admission) from a French multicenter prospective observational study: the AmarCAND2 cohort,1Leroy O. Bailly S. Gangneux J.P. Mira J.P. Devos P. Dupont H. et al.Systemic antifungal therapy for proven or suspected invasive candidiasis: the AmarCAND 2 study.Ann Intensive Care. 2016; 6: 2Crossref PubMed Scopus (65) Google Scholar was included. Patients with primarily documented invasive candidiasis, defined either with a positive direct examination or a positive culture and a subsequent Candida species identification at the time of SAT initiation, were excluded. Patients without primary PIC, i.e., for whom only a suspicion of IC led to the initiation of SAT, were included and clustered into two groups according to SPIC onset or not. More details on the reason for initiation of SAT in patients with suspected IC were provided in the article by Leroy et al.1Leroy O. Bailly S. Gangneux J.P. Mira J.P. Devos P. Dupont H. et al.Systemic antifungal therapy for proven or suspected invasive candidiasis: the AmarCAND 2 study.Ann Intensive Care. 2016; 6: 2Crossref PubMed Scopus (65) Google Scholar Univariate analyses were performed to select variables with a p-value threshold of about 0.25. Selected variables were introduced into multivariate hierarchical models to identify factors independently associated with SPIC. The final model with the smallest Akaike information criterion was chosen. Risk factors that were tested and not retained in the final models were: (1) baseline characteristics: diabetes mellitus, immunosuppression, other comorbidity (other than diabetes, immunosuppression, hemopathy and cancer), SAPS II score; (2) at SAT initiation: invasive mechanical ventilation, SOFA score, previous length of ICU stay, previous length of hospital stay, broad spectrum antibacterial therapy in the past ten days, corticosteroid therapy, total parenteral nutrition, red blood cell and platelet transfusion, procalcitonin, creatinine, reactive C-protein, and multifocal Candida spp. colonization. Two sub-group analyses were performed by considering the type of ICU admission (intra-abdominal surgery vs. other). A non-parsimonious p value of <0.10 was retained for analyses. Statistical analyses were performed by using SAS v9.4 (SAS Institute Inc., Cary, NC, USA.). A total of 544 patients were included in the analysis, including 112 (21%) patients with SPIC (candidaemia: 19 (17%), intra-abdominal IC: 64 (57%) and deep-seated IC: 29 (26%), mainly from sterile biopsy (N = 17 – 59%)). Secondary PIC were more frequently observed in centers with an infectious disease unit, and the proportion of SPIC was larger in surgery ICUs compared to medical ICUs. At ICU admission, the rate of SPIC was higher for patients admitted for an intra-abdominal surgery (28%) compared to those admitted for another reason (other surgery or medicine: 14%) (p <0.01); there was no other significant difference between both groups (Table 1). At SAT initiation, the duration of ICU stay was shorter for patients with SPIC; otherwise, there were no significant differences between both groups. The raw mortality was 26% in the SPIC group and 33% in the rest of the sample, but the difference was not significant (Chi square p-value = 0.11). The multivariate analysis confirmed that patients admitted to the ICU for intra-abdominal surgery were more than three times as likely to experience an SPIC as the patients admitted for a medical reason. Septic shock at SAT initiation decreased the probability for SPIC. We may speculate that the attending physicians more often started antifungal drugs in case of septic shock, leading to a more frequent overuse of antifungal treatment. Other known IC risk factors were tested and showed no significant association with SPIC (Table 2). When we considered only those patients who were admitted to the ICU for an intra-abdominal surgery, we found an inverse association between septic shock and final diagnosis of SPIC. When we consider the sub-group of patients who were not admitted for an intra-abdominal surgery, no risk factor was found to be associated with SPIC (Table 2). The hierarchical model confirmed the significant variability of the rate of SPIC between centers, not explained by fixed center-effect variables unmasked in previous studies (see Table E1 ESM).3Azoulay E. Dupont H. Tabah A. Lortholary O. Stahl J.P. Francais A. et al.Systemic antifungal therapy in critically ill patients without invasive fungal infection*.Crit Care Med. 2012; 40: 813-822Crossref PubMed Scopus (102) Google ScholarTable 1Comparaison of the center and patient characteristics according to the presence of secondarily proven invasive candidiasisDescriptive statistics (N(%))Absence of secondary PICN = 432Secondary PICN = 112P-valueCenter characteristics University hospital322 (74.5)93 (83)0.06 ICU type<.01 Medicine64 (14.8)19 (17) Surgery97 (22.5)40 (35.7) Polyvalent271 (62.7)53 (47.3) Hematology unit in the hospital385 (89.1)91 (81.3)0.02 Infectious disease unit in the hospital313 (72.5)96 (85.7)<.01 Oncology unit in the hospital360 (83.3)93 (83)0.94 Infectious disease adviser in the ICU**i.e. At least one ICU attending physician is also graduated in infectious diseases.405 (93.8)107 (95.5)0.47 Local protocol for antifungal administration236 (54.6)56 (50)0.38 Local protocol for antibacterial administration319 (73.8)79 (70.5)0.48 Detection of Candida colonization280 (64.8)64 (57.1)0.13Patient characteristics on ICU admission Age (median [IQR])63.8 [53.2; 73]63.1 [54.5; 72.3]0.72 Sex (Male)268 (62)69 (61.6)0.93 Previous duration of hospitalization stay (days, median [IQR])2 [0; 10]1 [0; 7]0.57 SAPSII score48 [36; 62]48 [38; 61]0.98 Diabetes mellitus82 (19)15 (13.4)0.17 Immunodepression74 (17.1)16 (14.3)0.47 Malignant disease115 (26.6)40 (35.7)0.06 Other comorbidity*Other comorbidities: myocardial infarction, congestive heart failure, peripheral venous disease, stroke, dementia, chronic obstructive pulmonary disease, peptic ulcer disease, mild or severe chronic kidney disease, hemiplegia, mild or severe chronic liver disease.236 (54.6)58 (51.8)0.59 Type of ICU admission<.01 Intra-abdominal surgery189 (43.8)74 (66.1) Other surgery51 (11.8)12 (10.7) Medicine192 (44.4)26 (23.2)Patient characteristics on SAT initiation Previous duration of ICU stay (days, median [IQR])4 [1; 12]1 [0; 8]<.01 Candida score†Candida score was computed on the basis of available data (severe sepsis, surgery admission, parenteral nutrition and the knowledge of multifocal colonization.3 [2; 3]2.5 [1; 3]0.20 SOFA score8 [5; 11]7 [5; 10]0.05 Septic shock268 (62)62 (55.4)0.20 Invasive mechanical ventilation366 (84.7)92 (82.1)0.50 Central venous catheter417 (96.5)109 (97.3)0.68 Urinary catheterization408 (94.4)108 (96.4)0.40 Hemodialysis or hemodiafiltration131 (30.3)29 (25.9)0.36 Total parenteral nutrition185 (42.8)50 (44.6)0.73 Antibacterial therapy384 (88.9)97 (86.6)0.97 Corticosteroid treatment124 (28.7)27 (24.1)0.33 Multifocal Candida spp colonization‡Multifocal Candida spp colonization corresponded to one of the possible reasons for initiating a treatment in a patient with suspected IC.116 (26.9)26 (23.2)0.43 Red blood cell transfusion232 (53.7)57 (50.9)0.60 Platelet transfusion94 (21.8)25 (22.3)0.90 C-reactive protein (mg/L)152 [83.2; 234]173 [102; 255]0.23 Creatinine (μmol/L)104.5 [65; 168]99.5 [56; 176]0.55 Procalcitonine (ng/L)2320 [400; 10000]3640 [360; 11490]0.54OR >1 in favor of secondary candidemia.; IQR: interquartile range.* Other comorbidities: myocardial infarction, congestive heart failure, peripheral venous disease, stroke, dementia, chronic obstructive pulmonary disease, peptic ulcer disease, mild or severe chronic kidney disease, hemiplegia, mild or severe chronic liver disease.** i.e. At least one ICU attending physician is also graduated in infectious diseases.† Candida score was computed on the basis of available data (severe sepsis, surgery admission, parenteral nutrition and the knowledge of multifocal colonization.‡ Multifocal Candida spp colonization corresponded to one of the possible reasons for initiating a treatment in a patient with suspected IC. Open table in a new tab Table 2Result of the multivariate hierarchical model for secondary candidiasisSecondary candidiasisN (%)Variables selected at the last step of the multivariate analysisNoYesOR [95% CI]P valueAll patientsN = 544432 (79)112 (21)Type of ICU admission<.01 Intra-abdominal surgery3.06 [1.69; 5.54] Other surgery1.88 [0.79; 4.48] MedicinerefSeptic shock0.62 [0.38; 1.03]0.06Malignant hemopathy and cancer1.59 [0.95; 2.63]0.08Patients with intra-abdominal surgery (N = 263)189 (72)74 (28)Septic shock0.53 [0.28; 0.98]0.04Malignant hemopathy and cancer1.70 [0.92; 3.12]0.09Patients without intra-abdominal surgery (N = 281)243 (86)38 (14)No factor associated with PICOR>1 in favor of a secondarily documented invasive candidiasis. Open table in a new tab OR >1 in favor of secondary candidemia.; IQR: interquartile range. OR>1 in favor of a secondarily documented invasive candidiasis. To conclude, the first observation was that in a large prospective cohort of patients who receive an empirical SAT, only 21% experienced a documented invasive candidiasis. Intra-abdominal surgery was identified as the main and sole risk factor allowing differentiation of patients with empirical SAT at risk for SPIC. Conversely, other conventional risk factors such as total parenteral nutrition, severe sepsis, and previous antibacterial therapy, which are part of usual clinical scores,2Leon C. Ostrosky-Zeichner L. Schuster M. What's new in the clinical and diagnostic management of invasive candidiasis in critically ill patients.Intensive Care Med. 2014; 40: 808-819Crossref PubMed Scopus (85) Google Scholar were unable to identify patients at risk for SPIC. However, the absence of data on multifocal Candida colonization, or on biological markers, was a study limitation. Considering that empirical SAT of non-transplant non-neutropenic patients has not demonstrated its efficiency,9Timsit J.F. Azoulay E. Schwebel C. Charles P.E. Cornet M. Souweine B. et al.Empirical Micafungin Treatment and Survival Without Invasive Fungal Infection in Adults With ICU-Acquired Sepsis, Candida Colonization, and Multiple Organ Failure: The EMPIRICUS Randomized Clinical Trial.JAMA. 2016; PubMed Google Scholar, 10Bailly S. Bouadma L. Azoulay E. Orgeas M.G. Adrie C. Souweine B. et al.Failure of empirical systemic antifungal therapy in mechanically ventilated critically ill patients.Am J Respir Crit Care Med. 2015; 191: 1139-1146Crossref PubMed Scopus (49) Google Scholar and that usual risk factors are clearly not sufficient to optimize adequacy of empirical antifungal therapy, new biomarkers or new diagnostic tests are urgently needed. Members of the independent Scientific Committee: Olivier Leroy (M.D., medical intensivist), Jean-François Timsit (M.D., PHD, medical intensivist, epidemiologist), Jean-Pierre Gangneux (MD, PhD, mycologist), Elie Azoulay (MD, PhD, medical intensivist), Jean-Michel Constantin (MD, PhD, anaesthesiologist and intensivist), Hervé Dupont (MD, PhD, anaesthesiologist and intensivist), Olivier Lortholary (MD, PhD, infectious diseases specialist), Jean-Paul Mira (MD, PhD, medical intensivist), Philippe Montravers (MD, PhD, anaesthesiologist and intensivist), Pierre-François Perrigault (MD, anaesthesiologist and intensivist) Manuscript preparation: the authors thank Celine Feger, M.D., (EMIBiotech), who provided assistance in preparing and editing the manuscript. Funding: This work was supported by MSD France, which was the sponsor. Conflicts of interest: OL has been consultant to Astellas, Gilead, Merck, Novartis, Pfizer and Sanofi. HD has been a consultant to Astellas, Gilead, Cubist, Astrazeneca, Merck and Pfizer. JFT has given lectures for symposiums set up by Astellas, Pfizer, MSD, 3M, Novartis, and Gilead; has benefited from unrestricted research grants to his research unit from 3 M, MSD, and Astellas; and has been a consultant involved in scientific boards for MSD, 3M, and Bayer. SB have no conflict of interest.