Cryptococcal antigen detection in broncho-alveolar lavage fluid

Medical Mycology, 2018, 56, 774–777 doi: 10.1093/mmy/myx092 Advance Access Publication Date: 26 October 2017 Brief Report Brief Report Cryptococcal antigen detection in broncho-alveolar lavage fluid Y. Senghor1 , J. Guitard1,2 , A. Angoulvant3,4 and C. Hennequin1,2,∗ AP-HP, Hôpital Saint-Antoine, Service de Parasitologie-Mycologie, F-75012, Paris, France, 2 Sorbonne Universités, UPMC Université Paris 06, Inserm UMR S 1135, CNRS ERL 8255, Centre d’Immunologie et des Maladies Infectieuses (CIMI-Paris), F-75013, Paris, France, 3 APHP, Unité de Parasitologie-Mycologie, Hôpital Bicêtre, F-94270, Kremlin Bicêtre, France and 4 Univ Paris Sud, GQE– Le Moulon, INRA – Université Paris-Sud – CNRS – AgroParisTech, CNRS, 91400 Orsay, France ∗ To whom correspondence should be addressed. C. Hennequin, MD, PhD, Head of the Department of Parasitology-Mycology, Hôpital St Antoine, Paris. E-mail: Received 8 May 2017; Revised 25 July 2017; Accepted 29 August 2017; Editorial Decision 9 August 2017 Abstract Cryptococcal antigen (CryAg) testing in serum and CSF is a clue diagnostic tool for cryptococcosis. In this study, we reviewed the performances of the CryAg detection (Premier EIA, Meridian) routinely performed in broncho-alveolar lavage fluid (BALF) during a 7year period (2007–2013). CryAg was detected in 12 cases among 4650 BALF analyzed, while positive culture from BALF was detected in nine cases. We found sensitivity, specificity, positive and negative predictive values at 0.44–0.80 (according to the radio-clinical form), 0.99, 0.36, and 0.99, respectively. These results do not support the routine use of the test in BALF. Key words: Cryptococcosis, pneumonia, diagnosis, antigen. Cryptococcosis is mainly acquired following inhalation of spores from the environment and is in fact originally pulmonary. Although this primo-infection seems to be mostly a- or pauci-symptomatic, it allows the fungus to remain quiescent in the alveoli before a possible reactivation due to immunosuppression, mainly induced by the human immunodeficiency virus (HIV) infection or immunosuppressive therapies.1,2 Neurological symptoms are then often predominant but pulmonary symptoms may sometimes be foreground, notably in HIV-negative patients.3,4 Mostly depending on the immune status of the patient, the clinical presentation and outcome of pulmonary cryptococcosis are highly variable, ranging from asymptomatic colonization to severe pneumonia with respiratory failure.5 In a consistent clinical context, a definitive diagnosis of pulmonary cryptococcosis relies on the positivity of cultures for 774 Cryptococcus neoformans from bronchopulmonary specimens, sometimes anticipated by the demonstration of encapsulated yeasts on direct examination (Indian ink test). However, this may be difficult to achieve as it requires microscopic expertise and because concurrent microorganism growth may mask the presence of C. neoformans in cultures. Indeed, encapsulated yeasts may be missed due to the cellularity of the broncho-alveolar lavage fluid (BALF) or, by the contrary, some cells may be mistaken for yeasts leading to false-positive results. The culture, if it turns positive, requires at least 48 to 72 hours of incubation. Due to these limitations and because the cryptococcal antigen (CryAg) detection in serum and cerebrospinal fluid (CSF) is associated with excellent performances, it was tantalizing to test CryAg in broncho-pulmonary specimens.6 To the best of our knowledge, there are only four reports  C The Author 2017. Published by Oxford University Press on behalf of The International Society for Human and Animal Mycology. All rights reserved. For permissions, please e-mail: 1 Senghor et al. anticipated the diagnosis of cryptococcosis by 3 days (case 2). All of these patients but one (case 5) had a positive CryAg result in their BALF, leading a sensitivity rate estimated at 0.80. Four patients were considered colonized (cases 6–9). All of these patients were negative for CryAg in BALF. One patient was immunocompromised. Two had underlying lung diseases (lung fibrosis and acute lung edema). It has already been reported that intercurrent pulmonary disease may favor the cryptococcal pulmonary colonization.11 It is possible that in such cases, the fungal load was low enough that the immunologic test remains negative. However, we cannot rule out the fact that this condition represents the first stage of an invasive infection, as it has been previously described.3 Such an outcome could have been stopped thanks to an early antifungal therapy initiated in three of these patients. Gathering active infection and colonization, sensitivity rate of testing CryAg in BALF drops at 0.44. Finally, seven patients (8 BALF) were considered to have a possible pulmonary cryptococcosis. One case corresponded to the follow-up a patient with definitive pulmonary cryptococcosis (case 4.2), and two patients suffered disseminated cryptococcosis (cases 10, 11). In these cases, one can hypothesize that the positivity of the BALF test could correspond to the effusion of the seric antigen which was positive at a high level in these patients. Finally, four patients had no other diagnostic tests positive for cryptococcosis (cases 12–15). In these cases, the culture was also negative for other basidiomycetous species such as Trichosporon or Rhodotorula, which are known to crossreact with the CryAg.12 None of these patients received antifungal therapy. One died the day after the BALF CryAg returned positive, preventing any definitive conclusion (case 15). Another died 20 days later due to a probable invasive aspergillosis (case 14). The other two survived more than 2 years, and none developed cryptococcosis. Taking these results together, four to seven results can be considered as false positive for the diagnosis of pulmonary cryptococcosis. Nevertheless, because the vast majority of patients without cryptococcosis had a negative BALF CryAg, the specificity of the test remained high at 0.99, while in contrast, the positive predictive value was lower, calculated at 0.36. Comparison of our results with those from previous studies is difficult since both the kit used and the definitions applied to classify the clinical forms differ (Table 2). However, with the exception of the study by Baughman et al., who used a personal cut-off for the kit they used, both the sensitivity and the positive predictive value appear insufficient. One can expect that the use of the more recently available immunochromatographic test would have enhanced the sensitivity.12,13 However, none of the manufacturers currently recommend the use of their kit for the focused on this utilization.7-10 However, most of these studies have been performed before the advent of highly active antiretroviral therapy and/or analyzed a limited number of specimens. Thus, to assess whether the CryAg detection in BALF is currently a valuable adjunctive test, we retrospectivel (...truncated)