Multicenter evaluation of a lateral-flow device test for diagnosing invasive pulmonary aspergillosis in ICU patients

Table 2 Aspergillus -specific lateral-flow device test performance for diagnosing probable or proven versus no invasive pulmonary aspergillosis^a

	Sensitivity	Specificity	PPV	NPV	DOR (95% CI)
Modified EORTC/MSG criteria [34]^b
Overall study population	80% (16/20)	81% (88/108)	44% (16/36)	96% (88/92)	17.6 (5.3 to 58.3)
Graz	83% (5/6)	79% (46/58)	29% (5/17)	98% (46/47)	19.2 (2.0 to 179.9)
Innsbruck	75% (6/8)	95% (35/37)	75% (6/8)	95% (35/37)	52.5 (6.2 to 447.6)
Vienna	80% (4/5)	54% (7/13)	40% (4/10)	88% (7/8)	4.7 (0.4 to 54)
Mannheim^c	100% (1/1)	78% (7/9)	33% (1/3)	100% (7/7)	9 (0.3 to 200)
Original revised EORTC/MSG criteria [34]
Overall study population	87% (13/15)	81% (101/124)	36% (13/36)	98% (101/103)	28.5 (6 to 135)
Clinical algorithm according to Blot et al. [35]
Overall study population	83% (10/12)	79% (108/137)	26% (10/39)	98% (108/110)	18.6 (3.9 to 89.7)

^aCI, Confidence interval; DOR, Diagnostic odds ratio; EORTC/MSG, European Organization for Research and Treatment of Cancer/Mycoses Study Group; NPV, Negative predictive value; PPV, Positive predictive value. ^bIncluding intensive care unit stay longer than 4 days as a newly introduced host factor and bronchoalveolar lavage fluid galactomannan >0.5. ^cResults from Mannheim represent Aspergillus-specific lateral-flow device test performance for probable or proven versus possible or no invasive pulmonary aspergillosis (IPA), owing to the fact that only proven and possible cases were reported by this center. Classification was carried out according to EORTC/MSG criteria with and without modifications (possible IPA cases were excluded), and proven or putative IPA versus no IPA was classified according to the clinical algorithm published by Blot and colleagues [35].

ISSN: 1364-8535