High False-Positive Rate of (1,3)-β-D-Glucan in Onco-Hematological Patients Receiving Immunoglobulins and Therapeutic Antibodies

Abstract

Immunoglobulins and/or therapeutic antibody preparations are associated with a high rate of false-positive (1,3)-β-D-glucan (BDG) tests in onco-hematological patients routinely screened for fungal infections. The benefit of BDG monitoring shall be balanced against the risk of false-positive tests leading to unnecessary investigations and costs in this population.

β-D-glucan, intravenous immunoglobulines, therapeutic antibodies, false-positive, onco-hematology

Invasive fungal infections (IFIs) are responsible for substantial morbidity and mortality in onco-hematological patients. While early IFI diagnosis is important to achieve a favorable outcome, it is often difficult to obtain without invasive procedures, which are associated with hemorrhagic and/or infectious complications in such patients. Hence, serum fungal markers such as galactomannan (GM) or (1,3)-β-D-glucan (BDG) have been increasingly used for the diagnosis of IFIs.

BDG is a polysaccharidic component of the fungal cell wall that can be detected in plasma by using a colorimetric test based on the detection of activated factor G from the horseshoe crab Limulus polyphemus amoebocyte lysate (LAL) [1]. The test can detect BDG from most pathogenic fungi, except for Mucorales, Cryptococcus spp., and Blastomyces dermatitis, which release only little BDG in serum. According to various meta-analyses, BDG sensitivity and specificity to detect IFIs range between 77–80% and 63–85%, respectively [2, 3]. False-positive testing may be due to the test’s interference with antibiotics, bacteremia, enterocolitis, cellulose dialysis filters, or surgical gauzes and blood-derived products such as intravenous immunoglobulins (IVIGs) [4, 5]. We conducted a comprehensive study to analyze the cause of false-positive BDG in a cohort of 241 onco-hematological patients undergoing 375 hospitalization courses.

METHODS

This study was conducted in a cohort of patients with hematological malignancies at the Lausanne University Hospital (Switzerland; Swissethics 2017-01975). This hematology unit is dedicated to patients with acute leukemia receiving induction or consolidation/maintenance chemotherapy, as well as patients with lymphoma or multiple myeloma undergoing autologous hematopoietic stem cell transplantation (auto-HSCT) or allogeneic HSCT (allo-HSCT). Of note, allo-HSCTs are performed in another hospital. All patients who had at least 1 BDG measurement from February 2017 (date at which the test was implemented) to August 2019 were included.

Patients with neutropenia expected to exceed 10 days (eg, induction or consolidation chemotherapy for acute leukemia and aplastic anemia) receive fluconazole prophylaxis to prevent invasive candidiasis and undergo simultaneous biweekly monitoring of serum BDG (Fungitell; Associates of Cape Cod, Inc) and GM (Platelia Aspergillus; Bio-Rad Laboratories) as part of a preemptive strategy for IFIs. Patients with neutropenia expected to be shorter than 10 days (auto-HSCT) do not receive antifungal prophylaxis and are tested with BDG and GM upon clinical suspicion. Patients with previous allo-HSCT under immunosuppressive drugs for the treatment of graft versus host disease receive posaconazole prophylaxis.

Neutropenic patients developing fever are treated with broad-spectrum antibiotics (cefepime, piperacillin-tazobactam, meropenem or imipenem-cilastatin). Diagnosis and classification (proven/probable/possible) of IFIs are established based on radiological and microbiological assessment according to the European Organization for Research and Treatment of Cancer and Mycoses Study Group criteria [6].

A BDG of 80 pg/mL or greater was considered true positive when obtained within an interval from 10 days before a diagnosis of a proven/probable/possible IFI until hospital discharge, and was considered false positive otherwise. Factors associated with false-positive BDG tests were analyzed by using a univariate model, considering a time window starting 14 days before the first positive BDG value and ending 5 days after the last positive BDG for each hospital course. The following conditions were considered potential causes of false-positive BDG testing: transfusions of blood products, IVIG (Kiovig [Takeda Pharma] or Privigen [CSL Behring]), antibiotics, bacteremia, hemodialysis, mucositis, neutropenic enterocolitis, enteral/parenteral nutrition, chemotherapeutic agents, monoclonal/polyclonal antibodies, tyrosine kinase inhibitors, and bortezomib. Statistical analyses were performed using STATA version 16 (StataCorp).

RESULTS

The study included 241 patients undergoing 375 hospital courses, among which 15 were characterized by a proven or probable IFI (4%). There were 147 males (61%) and a median age of 59 years (interquartile range [IQR], 16 years). Seventy-eight patients (32%) suffered from acute myeloid leukemia, 51 (21%) from non-Hodgkin lymphoma, 49 (20%) from multiple myeloma, 19 (8%) from myelodysplastic syndrome, 14 (6%) from acute lymphoblastic leukemia, 13 (5%) from Hodgkin lymphoma, and 17 (7%) from other hematological diseases (Supplementary Table 1). A total of 1728 BDG tests were performed (median, 4 per hospital stay; IQR, 1–7), among which 22 were true positive, 59 false positive, 1534 true negative, and 113 false negative, providing 16.3% sensitivity, 96.3% specificity, and 15.5% positive- and 96.5% negative-predictive values, respectively (Supplementary Table 2).

In order to detect factors associated with false-positive BDG tests, we compared drugs and conditions at risk during hospital stays with (n = 32) and without (n = 343) false-positive tests (Table 1). The administration of IVIG was significantly more frequent in the former compared with the latter (25% vs 4%; odds ratio [OR] = 7.8; 95% confidence interval [CI], 3.0–20; P < .0001). Although monoclonal or polyclonal antibodies were rarely administered, several of them were or tended to be more frequent during hospital stays with false-positive BDG tests compared with the other hospital stays without false-positive BDG, including anti-thymocyte globulins (P = .02), daratumumab (P = .07), inotuzumab (P = .08), and obinutuzumab (P = .09). When all these drugs were pooled together, their administration was significantly associated with false-positive BDG testing (22% vs 6%; OR = 4.7; 95% CI, 1.8–12; P = .001). In addition, neutropenic enterocolitis was more frequent during hospital stays with false-positive BDG testing compared with the other (34% vs 19%; OR = 2.3; 95% CI, 1.1–5.0; P = .04), as well with as enteral nutrition (9% vs 2%; OR = 4.3; 95% CI, 1.1–17.2; P = .04). While no antibiotic was significantly associated with false-positive BDG testing, there was a trend towards an association for cefepime (72% vs 54%; OR = 2.2; 95% CI, .98–4.8; P = .06).

Table 1.

Factors Associated With False-Positive (1,3)-β-D-Glucan Testing

Drugs or clinical conditions	False-Positive (n = 32)		Other (n = 343)		OR	95% CI	P
	n	F	n	F
Blood products
Intravenous immunoglobulins	8	0.25	14	0.04	7.8	(3.0–20)	<.0001
Platelets	26	0.81	227	0.66	2.2	(.89–5.5)	.09
Red blood cells	22	0.69	221	0.64	1.2	(.56–2.6)	.6
Frozen plasma	4	0.13	48	0.14	.9	(.29–2.6)	.8
Antibiotics
Cefepime	23	0.72	185	0.54	2.2	(.98–4.8)	.06
Piperacillin-tazobactam	19	0.59	164	0.48	1.6	(.76–3.3)	.2
Amoxicillin-clavulanate	4	0.13	28	0.08	1.6	(.53–4.9)	.4
Meropenem	4	0.13	38	0.11	1.2	(.38–3.4)	.8
Imipenem	5	0.16	61	0.18	.9	(.32–2.3)	.8
Vancomycin	6	0.19	89	0.26	.7	(.26–1.6)	.4
Monoclonal/polyclonal antibodies
Daratumumab	2	0.06	0	0.00	…	…	.07c
Rituximab	0	0.00	8	0.02	…	…	1.0c
Inotuzumab	2	0.06	5	0.01	4.5	(.84–24)	.08
Blinatumomab	0	0.00	5	0.01			1.0c
Obinutuzumab	1	0.03	0	0.00	…	…	.09c
Anti-thymocyte globulin (rabbit)	2	0.06	2	0.01	11.4	(1.6–83)	.02
Pooleda	7	0.22	19	0.06	4.7	(1.8–12)	.001
Kinase inhibitors
Ruxolitinib	1	0.03	13	0.04	.8	(.1–6.5)	.8
Sorafenib	0	0.00	4	0.01	…	…	1.0c
Ponatinib	0	0.00	1	0.00	…	…	1.0c
Chemotherapeutic agents
Cyclophosphamide	2	0.06	49	0.14	.4	(.09–1.7)	.2
Melphalan	7	0.22	87	0.25	.8	(0.3–2.0)	.7
Etoposide	10	0.31	108	0.31	1.0	(0.5–2.2)	1.0
Doxorubicin	0	0.00	2	0.01	…	…	1.0c
Azacitidine	0	0.00	14	0.04	…	…	.6c
Mercaptopurine	2	0.06	42	0.12	.5	(.1–2.1)	.3
Methotrexate	2	0.06	45	0.13	.4	(.1–1.9)	.3
Asparaginase	0	0.00	12	0.03	…	…	.6c
Tretinoin	1	0.03	5	0.01	2.2	(.25–19)	.5
Bortezomib	0	0.00	1	0.00	…	…	1.0c
Nutrition
Enteral	3	0.09	8	0.02	4.3	(1.1–17.2)	.04
Parenteral	8	0.25	62	0.18	1.5	(.7–3.5)	.3
Infectious complications
Bacteremia	8	0.25	100	0.29	.8	(.4–1.9)	.6
Neutropenic enterocolitis	11	0.34	64	0.19	2.3	(1.1–5.0)	.04
Mucositis
Absent	23	0.72	211	0.62	Ref
Grade I–IIb	5	0.16	70	0.20	.7	(.2–1.8)	.4
Grade III–IVb	4	0.13	62	0.18	.6	(.2–1.8)	.3

Abbreviations: CI, confidence interval; F, frequency; n, number of hospital stays; OR, odds ratio; Ref, reference.

One patient received 2 monoclonal antibodies.

Grades according to National Cancer Institute–Common Toxicity Criteria.

P value was calculated by using the Fisher’s exact test.

False-positive BDG tests were observed at a median of 6 days (IQR, 3–10 days) and 20 days (IQR, 12–27 days) following the administration of IVIG and monoclonal antibodies, respectively (Figure 1).

Figure 1.

Values of BDG tests before and after the administration of IVIGs (black circle, upper panel) or M/PABs (white circles, lower panel) in hospital courses with no IFIs. Dashed line indicates the BDG-positive cutoff at 80 pg/mL; the dotted line indicates the day of IVIG or M/PAB administration. Abbreviations: BDG, (1,3)-β-D-glucan; IFI, invasive fungal infection; IVIG, intravenous immunoglobulin; M/PAB, monoclonal/polyclonal antibody.

DISCUSSION

Since BDG has been reported as a reliable test for early detection of IFIs in onco-hematological patients, some centers, including ours, started to implement once- or twice-weekly routine BDG tests a part of preemptive strategies. While sparse cases of false-positive BDG testing are increasingly reported [7, 8], the impact of false-positivity has not been evaluated systematically in a clinical setting. Here, we analyzed the yield of twice-weekly routine BDG testing in 241 patients undergoing 375 hospital courses for the treatment of onco-hematological disease.

In this cohort, most patients with positive BDG tests did not have a fungal infection. Administration of IVIGs and monoclonal/polyclonal antibodies were the 2 most important causes of false-positive BDG tests. To our knowledge, this is the largest clinical report of false-positive BDG tests in an adult population, and the first one to show an association with monoclonal/polyclonal antibodies. Falsely elevated BDG resolved after 1 to 2 weeks in most cases.

Contamination of IVIG product with LAL-reactive material has been described since 1989 [7]. This material is supposedly released from the elution of cellulose membrane in depth filters used for blood clarification during IVIG manufacturing [7, 9]. Inconsistent data regarding BDG false-positivity may result from important differences in the local practice for IVIG administration. In some centers, IVIGs are used to compensate for the functional immune deficiency caused by intensive chemotherapy regimens that might persist for 6 or more months after the end of the treatment.

Our data suggest that monoclonal/polyclonal antibodies are also responsible for false-positive tests. Only a small number of patients received each of these specific medications, but a pooled analysis shows a significant association. To our knowledge, this finding was never previously reported and could be of particular importance in the future, as these medications are increasingly prescribed in the treatment of onco-hematological diseases. The mechanism responsible for false-positive BDG testing is unknown but could be linked to contamination with LAL-reactive material during the purification process through depth filters, as for IVIGs.

Few other factors were associated with false-positive BDG testing in our study. A significant association with neutropenic enterocolitis was observed, which might be due to intermittent translocation of fungal elements such as Candida species through inflamed digestive mucosa, albeit not demonstrated [10]. Enteral nutrition was also significantly associated with false-positive BDG tests. This could be explained by the presence of complex sugars contained in enteral feeding preparations that may cross-react with BDG testing [11]. There was no significant association with parenteral nutrition. The sole antibiotic that tended to be associated with false-positive BDG tests in this study was cefepime. The association of antibiotics with false-positive BDG tests seems to mainly result from in vitro studies, in which BDG is tested directly from highly concentrated antibiotic preparations (eg, vials), which may contain LAL-reactive material [2, 4, 5, 12]. However, such associations are rarely confirmed in clinical studies.

The limitations of this study include its retrospective design and the relatively small sample size and limited observation period, with a low incidence of IFIs. Furthermore, the cause of false-positive BDG tests remains unknown in a number of cases.

Altogether, IVIGs and monoclonal/polyclonal antibodies were significantly associated with false-positive BDG tests in a cohort of patients with hematologic malignancies. Together with the low frequency of IFIs, which partially rely on the use of antifungal prophylaxis, this observation contributed to a low yield of BDG testing. In conclusion, the benefit of BDG monitoring in onco-hematological patients should be balanced against the risk of false-positive tests, which may lead to unnecessary investigations and costs, especially when patients are receiving IVIGs or other antibody preparations.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. The authors thank Aurélie Guillet, Corine Guyat-Jacques, and Monika Ochsner for their important contribution in collecting clinical data as well as Fady Fares for data management. They also thank Thierry Calandra, Oscar Marchetti, and Jacques Bille for facilitating the implementation of routine BDG testing in the Isolation Unit.

Financial support. P.-Y. B. is supported by the Swiss National Science Foundation (grant numbers 31CA30_196036, 33IC30_179636, and 314730_192616), the Leenaards Foundation, the Santos-Suarez Foundation, as well as grants allocated by Carigest. A.-S. B. received a grant from Promex Stiftung für die Forschung (via Carigest SA).

Potentials conflicts of interest. F. L. reports research grants from Novartis; Merck, Sharp & Dohme; Pfizer; Swiss National Science Foundation; and Santos-Suarez Foundation outside of the submitted work; and consulting fees from Gilead for Advisory Boards. P.-Y. B. received consulting fees for advisory board from Pfizer and honoraria for a lecture and support to attend meetings from Gilead. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.