Investigation of discordant positive Sōna lateral flow assay results for detection of Coccidioides antibodies

LETTER

Coccidioidomycosis is a fungal infection caused by the dimorphic fungi Coccidioides immitis and Coccidioides posadasii (1). Coccidioides is found in the soils of dry, arid regions of the southwestern United States and the northern regions of Mexico. Most cases in the U.S. occur in Arizona, with about 30% of cases happening in California (2). The clinical presentation of coccidioidomycosis ranges from asymptomatic infection (~60% of individuals) to pulmonary and extrapulmonary disease (3, 4). Patients at highest risk of diseases are those who are immunocompromised, including, but not limited to, persons with HIV/AIDS (5), transplant recipients (6), pregnant individuals (7), and those with diabetes mellitus (8).

Diagnosis of coccidioidomycosis has historically relied on serologic testing (9) with immunodiffusion (ID) and complement fixation (CF), which are slow and labor-intensive (10). More recently, simpler and automated testing using enzyme immunoassays (EIA) has been introduced (11). Furthermore, a lateral flow assay (LFA; Sōna; IMMY, Norman, OK) was most recently approved for rapid testing. A study evaluating the Sōna LFA in Southern California reported a sensitivity of 92.0% (92/100) and a specificity of 99.3% (802/808) using EIA and ID as the reference methods (12). Similar performance was observed in a smaller validation study conducted at our institution in Northern California. Conversely, in a study from an endemic area, the Sōna LFA demonstrated 31% sensitivity compared to standard EIA (Clarus Coccidioides Ab Enzyme Immunoassay; IMMY) for the early diagnosis of coccidioidomycosis (13). In October 2021, the Stanford Health Care clinical microbiology laboratory started utilizing the Sōna LFA for clinical testing of serum from patients with suspected coccidioidomycosis and in asymptomatic patients undergoing solid organ transplant evaluation. Positive LFA samples were sent to the coccidioidomycosis serology laboratory at UC Davis for a CF test. In the first year of testing with LFA, the overall positivity rate was 4.9% (95/1930). However, we noticed discordant LFA-positive/CF-negative results in 70.1% (53/75) of LFA-positive samples collected from unique patients, raising concern for false-positive LFA results. To investigate discordant results, EIA by IMMY (Norman, OK) and Meridian Biosciences (Cincinnati, OH), detecting IgM and IgG, was performed on 36 discordant samples with sufficient volume. These belonged to unique patients, of which 21 (58.3%) lived in an endemic region. Symptomatic patients formed 55.6% (20/36), and 44.4% (16/36) were from asymptomatic pre-transplant patients. 20/36 (55.6%) were from immunocompromised patients. Additionally, we tested 18 LFA+/CF+ and 10 samples that were LFA-negative and not tested by CF (Table 1). This study was a quality improvement project, and, therefore, approval from the Stanford University institutional review board was waived.

TABLE 1.

Investigation of discordant patient serum samples tested for Coccidioides antibodies by Sōna lateral flow assay and complement fixation with two enzyme immunoassays

Group	No. of samples	Immy IgM		Meridian IgM		Immy IgG		Meridian IgG		Positive for Immy or Meridian EIA e (IgM or IgG)	Negative by Immy and Meridian EIA
		Positive	Negative	Positive	Negative	Positive	Negative	Positive	Negative
Positive control: f LFA+/CF+	18	5 (27.8%)	13 (72.2%)	6 (33.3%)	12 (66.7%)	17 (94.4%)	1 a (5.6%)	17 (94.4%)	1 a (5.6%)	18 (100%)	0 (0%)
Negative control: LFA−/CF d ND g	10	0 (0%)	10 (100%)	0 (0%)	10 (100%)	0 (0%)	10 (100%)	0 (0%)	10 (100%)	0 (0%)	10 (100%)
Discordant sample: LFA+/CF−	36	3 b (8.3%)	33 (91.7%)	5 (13.9%)	31 (86.1%)	2 c (5.6%)	34 (94.4%)	8 (22.2%)	28 (77.8%)	9 (25.0%)	27 (75.0%)

^a IgG negatives occurred in unique samples.

^b All three positives were positive for Meridian IgM.

^c Both positives were positive for Meridian IgG.

^d CF, complement fixation.

^e EIA, enzyme immunoassays.

^f LFA, lateral flow assay.

^g ND, not done.

For LFA-positive/CF-positive samples, 94.4% (17/18) were positive by both EIA; the two EIA were 88.9% (16/18) concordant. The 10 LFA-negative samples were 100% negative by both EIAs. For the 36 LFA-positive/CF-negative samples, three (8.3%) were IgM-positive by Immy and five (13.9%) were IgM-positive by Meridian; all three positives by Immy were also positive by Meridian. Testing for IgG revealed two (5.6%) positives by Immy and eight (22.2%) by Meridian. Both samples positive by Immy were also positive by Meridian. Of the eight IgG-positive patients, seven were symptomatic and one was asymptomatic. Our findings show the majority (77.8%) of LFA-positive/CF-negative samples were also EIA-negative, suggesting the positive LFA results likely represented false-positive results. The false-positive rate was significantly higher in asymptomatic patients compared to symptomatic patients [93.8% (15/16) vs 65.0% (13/20); P = 0.04]. However, 22.2% of LFA-positive/CF-negative samples were positive by at least one EIA, therefore suggesting true-positive results with LFA in this group.

To generate additional evidence for the conclusions derived from EIA results, we performed a chart review to obtain clinical outcomes and additional laboratory results and requested unreported ID results from the UC Davis serology laboratory. Of the 16 asymptomatic pre-transplant patients, two had positive ID. None went on to develop coccidioidomycosis, although eight were administered antifungal prophylaxis because of the positive LFA results and residence in an endemic region (3). Of the 20 symptomatic patients who were evaluated for coccidioidomycosis, seven had positive IgG from either Immy or Meridian (Table 2). Of these, four were diagnosed and treated for coccidioidomycosis based on a positive ID (n = 3) and Coccidioides PCR (n = 1). One patient with positive Coccidioides ID was suspected to have mild coccidioidomycosis, and the patient’s symptoms improved without treatment. One patient had a positive Histoplasma ID and a positive Coccidioides ID, which likely represented a cross-reaction of LFA with anti-Histoplasma antibodies, as is known to occur per the package insert. The last EIA-positive patient had a negative ID result and was not diagnosed with an infection. Among 13 EIA-negative patients, one patient with a positive ID and positive culture was diagnosed with coccidioidomycosis (Table 2). In summary, of seven patients with positive IgG by Meridian EIA, six (85.7%) were ID or PCR positive, supporting the reliability of the Meridian EIA to assess discordant LFA and CF results.

TABLE 2.

Profile of symptomatic patients with discordant results by Coccidioides antibodies by Sōna lateral flow assay and complement fixation

				Coccidioides serology
Study ID	Clinical presentation	Reside in endemic region	Immune suppression	Immy IgG	Meridian IgG	Immunodiffusion	MiraVIsta EIA (Source)	Coccidioides PCR (Source)	Fungal culture (Source)	Clinically diagnosed
012	Respiratory failure	No	Diabetes	POS c	POS	POS	ND e	NEG d (PL) h	NEG (CSF) g	Yes
024	Malaise, persistent of a left lung infiltrate, eosinophilia	Yes	No	NEG	POS	POS	ND	ND	NEG (TIS) i	Yes
031	Pulmonary fibrosis	Yes	Diffuse large B-cell lymphoma	NEG	POS	POS	ND	ND	NEG (BAL) f	Yes
039	Chronic productive cough	Yes	No	NEG	POS	NEG b	ND	POS (TIS)	NEG (TIS)	Yes
005	Persistent low-grade fever	No	No	NEG	POS	NEG b	NEG (PL)	ND	ND	No
029	Pulmonary nodules, dry cough, SOB	No	No	NEG	POS	POS b Histo ID POS a	ND	NEG (PL)	NEG (BAL)	No
038	Pulmonary nodules	Yes	No	POS	POS	POS	NEG	ND	NEG (SPU)	Yes
030	Chest pain, cough, SOB, fever, malaise	Yes	Diabetes	NEG	NEG	POS	ND	ND	POS (SPU)	Yes
004	Septic shock	No	Acute myeloid leukemia	NEG	NEG	NEG b	NEG (PL)	NEG (PL)	ND	No
010	Chronic cough, chest pain	No	No	NEG	NEG	NEG	ND	NEG (PL)	ND	No
011	Fever, neurological and GI complications	Yes	No	NEG	NEG	NEG b	NEG (CSF)	NEG (PL)	NEG (CSF)	No
013	Left pleural effusion	Yes	Heart transplant	NEG	NEG	NEG	ND	NEG (PL)	NEG (BAL)	No
015	Sepsis	No	No	NEG	NEG	NEG b	ND	ND	ND	No
016	SOB, non-productive cough	No	Lung transplant	NEG	NEG	NEG b	ND	NEG (PL)	NEG (BAL)	No
018	Chest pain, cough	No	Systematic lupus erythematosus	NEG	NEG	NEG b	NEG (PL)	NEG (PL)	ND	No
021	Fever	No	No	NEG	NEG	NEG b	NEG (CSF)	NEG (PL)	NEG (CSF)	No
025	Low-grade fever, night sweats	Yes	No	NEG	NEG	NEG b	NEG (PL)	ND	NEG (TIS)	No
027	Fever, slight cough	Yes	Lung adenocarcinoma	NEG	NEG	NEG b	ND	ND	ND	No
033	Back pain, weakness, eosinophilia	No	Kidney transplant	NEG	NEG	NEG b	NEG (PL)	ND	NEG (TIS)	No
034	Encephalitis, cognitive decline	No	Autoimmune encephalopathy	NEG	NEG	NEG b	ND	ND	ND	No

^a Also had a positive Histoplasmosis serology by immunodiffusion.

^b Test result was obtained for this study but was not available to providers for clinical decision-making.

^c POS, positive.

^d NEG, negative.

^e ND, not done.

^f BAL, bronchioalveolar lavage.

^g CSF, cerebrospinal fluid.

^h PL, plasma.

ⁱ TIS, tissue.

Based on the ID and EIA results, this study revealed that 66.7% (24/36) of Sōna LFA-positive/CF-negative results represented false-positive results. Our findings indicate further testing with ID or EIA, along with clinical judgment, is needed when LFA and CF results are discordant. Further studies with larger sample sizes and diverse geographic locations are needed to confirm our findings.