Bronchoalveolar Lavage Fungal Antigen Detection in Critically Ill Children

Abstract

Invasive fungal infections (IFI) remain a growing problem in immunocompromised individuals. Recent trends are related to improvements in medical care for a wide range of previously untreatable conditions. More patients are now treated with chemotherapy, broad spectrum antibiotics, parenteral nutrition, glucocorticoids, immunomodulators, and prosthetic devices.1–4 These factors may predispose patients to opportunistic colonization and subsequent infection. In immunocompromised children, mortality rates from IFI with pulmonary involvement have been reported to be as high as 67%.5,6 Critically ill children in pediatric intensive care units are an especially vulnerable group as illustrated in the featured study by Fattah et al.7 Of 35 patients who were prospectively followed, 27(77%) died; all were diagnosed with highly probable pulmonary IFI. Specific potentially life-saving therapy to combat IFI requires early and accurate diagnosis. However, fungal infections present a diagnostic challenge as demonstrated by Fattah et al.7 IFI is often presumptively diagnosed in a susceptible patient who fails to respond to broad spectrum antibiotics. In this study, all children, regardless of immune status, who spent more than 48 hours in the pediatric intensive care units were eligible for enrollment, whereas those who developed pulmonary infiltrates or unexplained persistent fever despite antibacterial therapy underwent diagnostic bronchoscopy. Although none of the study patients received antifungals before bronchoscopy, empiric treatment is frequently started in critically ill patients where clinical improvement often serves as confirmation of the diagnosis. This practice can be life-saving but may lead to increased microbial resistance, toxicity to the patient, and unnecessary costs.4,8 Until recently, a lack of consensus on definitions and diagnostic criteria made establishing a definitive diagnosis of IFI difficult. In 2002, the European Organization for Research and Treatment of Cancer and the National Institute of Allergy and Infectious Diseases Mycoses Study Group at the National Institutes of Health published diagnostic IFI guidelines for patients with cancer and hematopoietic stem cell transplants.9 A proven diagnosis requires histopathologic confirmation or a positive culture from a sterile site. In critically ill children, obtaining clinical specimens often requires invasive procedures, many of which may be contraindicated or technically difficult. The dependence on cultures from blood and other body fluids present additional challenges as they are less reliable with sensitivities of 50% to 70%.2–5,10 Novel methods of acquiring specimens and detecting fungi are, therefore, the focus of ongoing research, as highlighted by Fattah et al.7 In this study, the authors assessed the safety and utility of bronchoscopy with bronchoalveolar lavage (BAL) in establishing a definitive diagnosis of pulmonary IFI. The data presented by Fattah in a population of critically ill children concur with prior studies that reported only minor reversible complications after pediatric bronchoscopy.11,12 The diagnostic yield of routine BAL analysis has been reported to range from 27% to 84%, with the wide degree of variation attributed to differences in technique, immune status, and the concurrent use of antimicrobials.11–14 In the featured study, histologic examination did not reveal fungi, but BAL cultures were positive in 15/35 (43%) of patients. Candida was isolated in 12 (34%) patients, both Candida and Aspergillus in 2 patients (6%), and Aspergillus in 1 patient (3%). In addition to BAL culture and histology, the authors employed serologic assays for the detection of antigens to Aspergillus spp. and Candida spp. The platelia Aspergillus enzyme linked immunosorbent assay is a Food and Drug Administration approved test for the diagnosis of invasive aspergillosis (IA).2 It is based on the detection of galactomannan (GM), a highly immunogenic cell wall polysaccharide of the Aspergillus spp.. GM detection has high specificity but variable sensitivity with values from 29% to 88% depending on the site of infection, type of body fluid tested, and patient immune status.2,14,15 When positive, it precedes diagnosis by culture by 2 to 10 days and is considered a major criteria in establishing a probable diagnosis of IFI.2,9 GM detection in BAL has a sensitivity of 60% to 100% and a specificity of 90% to 98%, depending on the optical density index used.10,14–16 In the featured study, Fattah et al7 found 4 patients (11%) positive for Aspergillus on BAL specimens and 3 patients (9%) on serum testing. The interpretation of these results is difficult as the authors do not report antimicrobial exposure, specifically piperacillin-tazobactam, which has been linked to false positives.17 For invasive candida infections, the cell wall polysaccharides, mannan, and 1,3 β-glucan, are major markers used as adjuncts in diagnosis. The recommended tests for mannan are Platelia Candida Antigen and Platelia Candida Antibody.2 The serum sensitivity of the Platelia test seems to be higher for Candida albicans, Candida glabrata, and Candida tropicalis (80% to 100%), than for the detection of Candida krusei and Candida parapsilosis (40% to 50%).2,18 This is important as most neonatal invasive Candida infections are caused by C. parapsilosis.3 Like GM detection in IA, serum mannan precedes positive cultures in patients with invasive candidiasis by 2 to 15 days.18 The utility of mannan detection in other body fluids like BAL has not been well-described and Fattah et al7 are among the first to investigate this application. In their study, 22 (62%) of BAL specimens and 19 (54%) of serum samples were mannan positive. One of the dilemmas in interpreting BAL results is distinguishing invasive pathogens from contamination or colonization. A recent study by Clancy et al10 reported that GM detection in BAL fluid was highly sensitive in detecting molds, but was not helpful in distinguishing IA from colonization. In contrast, others have reported that BAL candida antigen detection was highly sensitive for Candida pneumonia and reliably excluded contamination and colonization using serum Candida antigen detection as the marker for significant invasive disease.19 In the featured study, delineation between invasive disease and colonization cannot be accurately determined. An autopsy with histopathology of lung tissue would have been valuable to prove infection and would have assisted with calculating the sensitivities and specificities of the tests used. Furthermore, a correlation between antigen positivity and immune status would have shed light on the diagnostic yield in neutropenic versus non-neutropenic patients. However, Fattah et al7 have shown that BAL fluid is superior to serum for fungal detection in pulmonary IFI and that antigen testing is more sensitive than routine culture and histology. The recent advances in polymerase chain reaction and antigen detection have expanded the role of bronchoscopy with BAL in critically ill patients. These methods, combined with risk stratification, add to the sensitivity and specificity of BAL and allow for earlier diagnosis of IFI.2,15 Fattah et al7 have demonstrated that bronchoscopy is safe to perform in children and that BAL has a significant diagnostic yield. These findings are consistent with the growing body of evidence, which supports the use of bronchoscopy with BAL in the early diagnosis of IFI. However, the sensitivity and specificity of diagnostic testing and the distinction between invasive disease and colonization require further investigation.