Comparison of a Novel Rapid Lateral Flow Assay to Enzyme Immunoassay Results for Early Diagnosis of Coccidioidomycosis

Abstract

Background

Coccidioidomycosis (CM) is a common cause of community-acquired pneumonia where CM is endemic. Manifestations include self-limited pulmonary infection, chronic fibrocavitary pulmonary disease, and disseminated coccidioidomycosis. Most infections are identified by serological assays including enzyme-linked immunoassay (EIA), complement fixation, and immunodiffusion. These are time-consuming and take days to result, impeding early diagnosis. A new lateral flow assay (LFA; Sōna; IMMY, Norman, OK) improves time-to-result to 1 hour.

Methods

We prospectively enrolled 392 patients with suspected CM, compared the LFA with standard EIA and included procalcitonin evaluation.

Results

Compared with standard EIA, LFA demonstrates 31% sensitivity (95% confidence interval [CI], 20–44%) and 92% specificity (95% CI, 88–95%). Acute pulmonary disease (74%) was the most common clinical syndrome. Hospitalized patients constituted 75% of subjects, and compared with outpatients, they more frequently had ≥3 previous healthcare facility visits (P = .05), received antibacterials (P < .01), and had >3 antibacterial courses (P < .01). Procalcitonin (PCT) was <0.25 ng/mL in 52 (83%) EIA-positive patients, suggesting infection was not bacterial.

Conclusions

When CM is a possible diagnosis, LFA identified nearly one-third of EIA-positive infections. Combined with PCT <0.25 ng/mL, LFA could reduce unnecessary antibacterial use by 77%.

valley fever, coccidioidomycosis, lateral flow assay (LFA), enzyme immunoassay (EIA), procalcitonin (PCT)

Coccidioidomycosis (CM), also known as San Joaquin Valley fever, is an invasive fungal infection endemic to Arizona, California, and Mexico, as well as parts of Central and South America [1]. The causative agents of CM are the dimorphic fungal species Coccidioides immitis and Coccidioides posadasii [2, 3]. Of all coccidioidal infections, 60% are asymptomatic. Most of the others present as community-acquired pneumonia (CAP) with symptoms such as cough, pleuritic chest pain, fever, chills, skin rash, arthralgia, and fatigue [4, 5]. As a result, until correctly diagnosed, patients with CM are likely to receive unnecessary antibacterial drugs, laboratory tests, imaging, and invasive procedures. All could contribute to delays in appropriate management of patients with CM, adverse health consequences, and additional costs [6–10].

Serology has been the primary modality for diagnosing CM [11, 12]. Enzyme-linked immunoassay (EIA) commercial kits have been widely used and recommended for screening patients with suspected CM. Positive test results strongly suggest CM [13, 14]. Other serologic tests, immunodiffusion (IMDF) and complement fixation (CF) antibody tests, are more labor intensive, complex, slower, and generally less sensitive than EIA [15]. For most clinicians, all of these tests are only available from reference laboratories, and results often do not become available for days or even weeks after specimen submission. These limitations present challenges for rapid and accurate CM diagnosis.

A lateral flow assay (LFA; Sōna; IMMY, Norman, OK) has received Food and Drug Administration approval to be performed in Clinical Laboratory Improvement Amendments (CLIA)–approved laboratories [16]. It is a dipstick assay that can be completed within 1 hour, even in small CLIA-approved laboratories. Reporting time is comparable to other rapid tests performed for emergency departments, urgent care facilities, or clinics in close proximity to a CLIA-approved laboratory. Here we report a prospective, observational study of patients with possible CM conducted at our institutions comparing this novel LFA test with a standard EIA test, both performed by personnel in CLIA-approved commercial laboratories. As a prospective study, it also describes characteristics of a population within a CM-endemic region whom clinicians suspect of having this infection.

METHODS

Study Population

Our prospective observational study was conducted from 2 January 2019 to 31 December 2019 at 2 southern Arizona academic medical centers Banner-University Medical Center Tucson (B-UMCT), Banner-University Medical Center Phoenix (B-UMCP) and outpatient clinics. The study was approved by the University of Arizona institutional review board (project number 1811085933A011). Prior to the study period, we organized educational sessions with emergency department and clinic personnel to both promote CM awareness and the study. During the study, the research team provided periodic one-to-one promotion with healthcare providers. Our research coordinators were alerted about potential subjects (1) directly via the Cerner electronic medical record (EMR) [17], (2) when emergency department clinicians ordered the CM EIA screening test, or (3) by outpatient clinicians. Outpatient subjects (n = 99) were contacted by a research coordinator and came to our research facility to participate. Patients were excluded if younger than 18 years old or had a prior CM history. Consenting subjects allowed study personnel access to their EMR, completed a medical questionnaire, and provided an additional blood sample for C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), procalcitonin (PCT), and the LFA test [16]. Because both emergency department and outpatient settings involved a “single” visit, there were no “dropouts.” EIA test kits from a single manufacturer (Clarus Coccidioides Ab Enzyme Immunoassay; IMMY) and a single laboratory (Sonora Quest Laboratories, Tempe, AZ) were utilized for screening serology in enrolled subjects. Typically, positive EIA results reflexed to IMDF tests and, if either the immunoglobulin (Ig) M or the IgG IMDF was positive, also to CF antibody titrations. Some clinicians ordered the IMDF for screening or both EIA and IMDF tests together, which was beyond our study’s control.

Data Collection

A research coordinator, a physician masters of public health student, and a medical student compiled data from the patient questionnaire including demographics (age, sex, ethnicity, race, and length of endemic residence), length of illness prior to admission, symptoms duration, antibacterial prescription history, and the patient’s perception of Valley fever. The EMR data including various signs and symptoms, immunocompromised state, and radiographic and microbiologic results were collected. EIA IgM and IgG results (whether negative, indeterminate, or positive) were recorded according to the manufacturer’s instructions. The actual EIA optical density (OD) results were also available for analysis. When available, IMDF, IgM, and IgG results and CF antibody titers were recorded.

The LFA testing was performed at B-UMCT laboratory by certified laboratory technologists using kits generously provided by the manufacturer. Shortly before the study began, company representatives trained the technologists in specimen preparation, testing, and the reporting of results. Each result was independently reported using our laboratory’s standard-of-care testing protocol by laboratory technologists and disagreements over band intensity results were resolved by B-UMCT’s director of microbiology laboratory, our study co-investigator (W. D. L.). These results were recorded in the hospital laboratory information system. Data were extracted from this system per protocol and stored using Research Electronic Data Capture (REDCap) tools hosted at the University of Arizona [18, 19].

Other tests (CRP, ESR, and PCT) were conducted utilizing standard protocols at the B-UMCT or B-UMCP.

Patient Classification

Patients were defined as having CM if either an EIA IgM or IgG test was positive. Since indeterminate EIA results were not “positive” they were interpreted as negative. Bands developing on LFA strips were recorded as 1+, 2+, 3+, and 4+ intensities for subanalyses. A 1+ intensity band or greater was considered positive.

Statistical Analyses

Laboratory and clinical results exhibited non–normal distributions and were assessed via nonparametric methods. Groups were compared with the chi-square test for categorical variables [20]. For continuous variables, the 2 sided Mann-Whitney U test was used for 2 independent samples, and Kruskal-Wallis tests were utilized for multiple samples [21, 22]. Statistical analyses were conducted using R version 3.6.0 (R Foundation for Statistical Computing) [23]. P values less than .05 were considered significant, with no correction for multiple comparisons.

RESULTS

Demographic and Clinical Characteristics of Enrolled Subjects

In this study, 402 patients consented to participate. Ten patients missing either a confirmed medical record number or a CM screening test result were excluded. Of the 392 remaining subjects, 75% were hospitalized (Table 1). On average, hospitalized subjects were slightly older compared with subjects recruited in the emergency departments or outpatient clinics. Most patients (91%) had resided in Arizona for more than 1 year and 84% were aware of Valley fever before hospitalization. Of note, more outpatient (35%) than hospitalized (14%) subjects thought their symptoms were secondary to Valley fever.

Table 1.

 Patients’ Demographic Characteristics

	Hospitalized (n = 293)		Outpatient a (n = 99)		Totalb (N = 392)
	n	(%)	n	(%)	n	(%)	P
Mean age (range), years	55(18–98)		50(19–93)		54(18–98)
Sex							.77
Male	141	(48.1)	49	(50.5)	190	(48.7)
Female	152	(51.9)	48	(49.5)	200	(51.3)
Ethnicity							.98
Hispanic	80	(27.8)	26	(28.6)	106	(28.0)
Non-Hispanic	208	(72.2)	65	(71.4)	273	(72.0)
Race							.29
African American	20	(7.0)	6	(6.3)	26	(6.8)
AI/AN	11	(3.8)	6	(6.3)	17	(4.5)
Asian	4	(1.4)	3	(3.2)	7	(1.8)
Caucasian	237	(82.9)	79	(83.2)	316	(82.9)
Other	14	(4.9)	1	(1.1)	15	(3.9)
Arizona resident							1.0
Yes	265	(93.0)	86	(92.5)	351	(92.9)
No	20	(7.0)	7	(7.5)	27	(7.1)
Length of residence							.09
6–12 months	11	(4.0)	9	(9.9)	20	(5.4)
More than 1 year	255	(92.4)	79	(86.8)	334	(91)
Nonresident	10	(3.6)	3	(3.3)	13	(3.5)
Patient knowledge of VF							.35
Yes	243	(85.6)	76	(80.9)	319	(84.4)
No	41	(14.4)	18	(19.1)	59	(15.6)
Patient belief of VF infectionc							<.01
Yes	37	(14.2)	31	(34.8)	68	(19.4)
No	184	(70.5)	36	(40.4)	220	(62.9)
Unknown	40	(15.3)	22	(24.7)	62	(17.7)

Abbreviations: AI/AN, American Indian/Alaskan Native; VF, Valley fever.

^aOutpatient admission status includes observational patients (n = 19).

^bA total of 402 patients were enrolled; 10 patients were excluded because of no admission data or missing medical record numbers; variable may not sum to admission status or total due to incomplete demographic responses.

^cPatient believed that his/her symptoms were caused by VF.

As shown in Table 2, the most commonly seen clinical syndrome was acute pulmonary manifestations (74%). Other presenting signs and symptoms including headache, skin, and other extrapulmonary complaints were present in 21% of participants. Fibrocavitary lesions and pulmonary nodules were found on chest imaging in 4.2 % (n = 16) and 1.3% (n = 5) of subjects, respectively. Immunocompromised patients were more likely to be hospitalized (P < .001). Also shown in the Table 2, more hospitalized patients had ≥3 previous visits to healthcare facilities for their symptoms (P = .05), received antibacterial medications (P < .01), and received >3 antibacterial courses (P < .01) compared with outpatients.

Table 2.

 Participants’ Clinical Patterns

	Hospitalizeda (n = 293)		Outpatientb (n = 99)		Total (N = 392)
	n	(%)	n	(%)	n	(%)	P
Clinical syndrome							.31
Acute pulmonary disease	216	(73.7)	74	(74.7)	290	(74.0)
Chronic/fibrocavitary diseasec	15	(5.1)	1	(1.0)	16	(4.2)
Asymptomatic pulmonary nodulec	4	(1.4)	1	(1.0)	5	(1.3)
Other presentationd	58	(19.8)	23	(23.2)	81	(20.7)
Number of previous visits							.37
0	110	(37.5)	40	(40.4)	150	(38.3)
1	64	(21.8)	24	(24.2)	88	(22.4)
2	32	(10.9)	15	(15.2)	47	(12.0)
3	28	(9.6)	8	(8.1)	36	(9.2)
>3	59	(20.1)	12	(12.1)	71	(18.1)
Antibiotic prescribed							<.01
Yes	229	(80.9)	47	(50.0)	276	(73.2)
No	54	(19.1)	47	(50.0)	101	(26.8)
Number of antibiotics prescribed							<.01
0	106	(36.2)	55	(55.6)	161	(41.1)
1	35	(11.9)	22	(22.2)	57	(14.5)
2	34	(11.6)	12	(12.1)	46	(11.7)
3	16	(5.5)	3	(3.0)	19	(4.8)
>3	102	(34.8)	7	(7.1)	109	(27.8)
Immunocompromised							<.01
Yes	171	(58.8)	27	(27.6)	198	(50.9)
No	120	(41.2)	71	(72.4)	191	(49.1)
Symptom duratione (IQR), days	14(54.0)		14(41.0)		14(40.0)		.60

Abbreviation: IQR, interquartile range.

^aA total of 402 patients were enrolled; 10 patients were excluded because of no admission data or missing medical record numbers; variable may not sum to admission status or total due to incomplete demographic responses.

^bOutpatient sum includes observational patients (n = 19).

^cChronic/fibrocavitary disease and pulmonary nodules assessed by imaging (n = 377).

^dOther presentation included nonpulmonary presentation (ie, fatigue, fever, muscle aches, night sweats, rash; n = 71), disseminated disease (ie, meningitis, skin lesion, prevertebral abscess; n = 10).

^eTotal days of previously experienced illness reported by patient upon consent.

Clinical presentations in the hospitalized and outpatient setting were similar as demonstrated in Figure 1. Most patients in both settings complained of fatigue, cough, and shortness of breath and these symptoms along with body aches were of the longest duration.

Figure 1.

 Median symptoms duration. ^aPercentages of patients reporting symptoms are displayed individually to the right of the bars. Abbreviation: SOB, shortness of breath.

Performance of the LFA Compared With the EIA and Other Serologies

The LFA tests were available with EIA for 370 subjects (Figure 2). Of the 65 subjects who were EIA positive (CM diagnosis rate, 17.6%) (Table 3), only 20 (31%) had positive LFA tests. On the other hand, of the 305 who were EIA negative, 280 (91.8%) were also LFA negative. Of these, 47 patients had indeterminate EIA results and only 2 (4%) were LFA positive. Also shown in Figure 2 is the distribution of LFA and IMDF results for 111 subjects, of whom 54 (49%) were tested because the EIA test was positive (data not shown). Of the 27 patients with positive IMDF tests, 11 (41%) had positive LFA tests. The overall agreement between LFA and both EIA and IMDF results are shown in Table 3.

Table 3.

 Novel LFA Agreement With EIA and Immunodiffusion

	EIA, %		IMDF, %
	Value	95% CI	Value	95% CI
Sensitivity,a %	30.8	19.9–43.5	40.7	22.4–61.2
Specificity,b %	92.0	88.1–94.6	95.2	88.3–98.7
PPV, %	44.4	32.2–57.5	73.3	48.8–88.8
NPV, %	86.2	84.1–88.0	83.3	78.5–87.3
PLR	3.8	2.2–6.5	8.6	3.0–24.7
NLR	0.8	.6–.9	0.6	.5–.9
Disease prevalence, %	17.57	13.8–21.8	24.3	16.7–33.4

PPV, NPV, PLR, NLR, and disease prevalence were calculated according to a standard formula using values shown in Figure 2.

Abbreviations: CI, confidence interval; EIA, enzyme-linked immunoassay; IMDF, immunodiffusion; LFA, lateral flow assay; NLR, negative likelihood ratio; NPV, negative-predictive value; PLR, positive likelihood ratio; PPV, positive predictive value.

^aSensitivity = true positive / (true positive + false negative) [eg, EIA sensitivity = 20/(20+45); Figure 2].

^bSpecificity = true negative / (true negative + false positive) [eg, EIA specificity = 280/(280+25); Figure 2].

Figure 2.

Novel LFA comparison against EIA and IMDF diagnostic tests. Abbreviations: EIA, enzyme-linked immunoassay; IMDF, immunodiffusion; LFA, lateral flow assay.

Because the sensitivity of the LFA in this study was lower than that in 2 previous studies [24, 25], we conducted subanalyses to identify possible explanations. In considering the duration of illness prior to study enrollment, subjects who were EIA positive/LFA positive tended to have been ill longer (median of 21 days) than those who were EIA positive/LFA negative (median of 14 days) (P = .17). Also, 7 of the 16 subjects who presented with fibrocavitary pulmonary findings were diagnosed with CM. Of these, 4 had a positive EIA and also a positive LFA. In the other 3 subjects, CM was diagnosed by culture, and both their EIA and LFA were negative. When OD values for EIA IgG and IgM in all subjects with positive EIA results were compared with the intensity of positive LFA results, no quantitative correlation was found (Supplementary Table 1). However, of EIA-positive subjects, the sum of the IgM and IgG OD values was significantly higher for subjects whose results were LFA positive compared with those who were LFA negative (P = .003). Finally, CF titers were performed for 17 of the LFA-positive and 46 of the LFA-negative subjects (Supplementary Table 2). Although the LFA band intensity did not appear to correlate with CF titers, LFA-negative results were significantly associated with a CF titer of less than 1:2 (P = .005). Similarly, 96% (44 of 46) of LFA-negative subjects compared with 65% (11 of 17) of LFA-positive subjects had CF titers less or equal to 1:4.

Proinflammatory Markers in EIA-Positive Subjects

The ESR, CRP, and PCT results for subjects with different EIA and LFA results are shown in Table 4. For EIA-positive patients, ESR values were elevated (>21 mm/hour for males, >31 for females) in 33 (52%), while CRP was greater than 8 ng/L in 43 (68%). Procalcitonin results were less than 0.25 ng/mL (bacterial infection unlikely [26]) in 52 of 63 patients (83 %) who were EIA positive. For these 52 patients, antibacterial status was available in 44 and 34 (77%) received 1 or more antibacterials. This was slightly higher than the 183 EIA-negative subjects with a PCT less than 0.25 ηg/mL and antibacterial usage was recorded where 133 (73%) had received antibacterial treatment. Subjects who were EIA negative/LFA negative (n = 272) had higher PCT results (median = 0.1 ng/mL, interquartile range = 0.5 ng/mL), suggesting an increased tendency towards bacterial infections (Supplementary Table 3). Similarly, although EIA-positive subjects frequently had elevated ESR and CRP, these values were higher in EIA-negative subjects. Other routine laboratory results are shown in Supplementary Table 3.

Table 4.

 Laboratory Result Thresholds for EIA and LFA Categories

	EIA+/LFA+ (n = 20)	EIA+/LFA− (n = 43)a	EIA−/LFA+ (n = 24)a	EIA−/LFA− (n = 276)a	P
PCT (ng/mL)					.01
<0.25	16	36	19	169
≥0.25	4	8	6	110
ESR (mm/hour)b					.04
<31.0 or 21.0	10	21	9	82
≥31.0 or 21.0	10	23	16	195
CRP (mg/L)					.01
<8.0	6	14	4	39
≥8.0	14	29	20	237

^aColumn sum reflects the lowest number of laboratory measurements available and may be lower than the total assays available: EIA+/LFA− (n = 45 – 2), EIA−/LFA+ (n = 25 – 1), EIA–/LFA− (n = 280 – 4).

Abbreviations: CRP, C-reactive protein; EIA, enzyme-linked immunoassay; ESR, erythrocyte sedimentation rate; LFA, lateral flow assay; PCT, procalcitonin; +, positive; −, negative.

^bReference range of 21 mm/hour was used for males and 31 mm/hour for females.

DISCUSSION

In this study, while the LFA demonstrated a high degree of specificity, its sensitivity compared with EIA results was only 31% (Table 3). The sensitivity of the LFA in relation to IMDF test results was slightly better but was still relatively low. As such, our results differ substantially from 2 prior studies demonstrating high sensitivity (non–peer- reviewed poster presentations) [24, 25].

False-negative EIA and IMDF tests occur frequently early in the course of coccidioidal infection, and so it is possible that some of the subjects with negative EIA or IMDF where the LFA test was positive might actually have a coccidioidal infection. However, the sensitivity of EIA and IMDF tests increases later in the course of infections [27], and it is also possible that a similar pattern might be true for the LFA. Since the design of our study was to enroll subjects as early as possible after the onset of illness, we considered this factor might be responsible for our different results. In fact, for EIA-positive subjects, those with a positive LFA test tended to have a longer duration of symptoms. Fibrocavitary pulmonary CM also implies a longer course of infection, and in all 4 subjects who were positive for both EIA and LFA that was the case. In addition, the OD values of positive EIA IgM results were significantly higher in LFA-positive than in LFA-negative subjects (P = .04). Similarly, the ESR and CRP levels were higher in subjects with a positive LFA. Finally, where CF results were available, they were more frequently low in LFA-negative subjects compared with those where the LFA was positive. Taken together, these observations suggest that the LFA was more sensitive in subjects with more protracted or extensive illnesses.

Whether this explains the discrepancy between our findings and those reported by others is not certain since the prior studies were conducted without clinical information related to the specimens tested. Both prior studies analyzed sera banked in reference laboratories. There is no information available from the earlier studies regarding whether specimens came from patients early in the course of their illness, whether the illness was uncomplicated pneumonia, or whether any of the patients had been receiving antifungal treatment. A previous report suggests that fluconazole treatment of primary coccidioidal infection reduces the likelihood of patients with CM developing positive IgG responses [28]. In contrast, by conducting a prospective enrollment of consenting subjects, we have much of this information available. Since we excluded patients with previously diagnosed CM, our cohort had a relatively recent onset.

In our study, 83% of EIA-positive subjects had PCT levels less than 0.25. Where antibacterial use was recorded, 77% of EIA-positive subjects with PCT level less than 0.25 received 1 or more antibacterial medications. Serum PCT assay has emerged as a biomarker to distinguish between bacterial and viral respiratory tract infections [26]. A prior study showed PCT to be low in patients with CM but noted that some of the measurements were taken some time after the initial CM diagnosis was made and 70% of the patients had already begun antifungal therapy [29]. Our study corroborates the conclusions of that prior study and supports the value of PCT in the differentiation of CM from bacterial CAP.

Our study also demonstrates an increased number of visits to healthcare facilities and utilization of antibacterial medications in hospitalized patients compared with subjects seen in the emergency department and outpatient clinics. This study reinforces 3 recent publications of delays in CM diagnosis, which led to unnecessary antibacterial use and increased healthcare costs. In a 2019 retrospective study, a delayed CM diagnosis occurred in 89% of patients (median delay = 23 days). In patients with delays, 1103 antibacterial orders, 22% of which were vancomycin and daptomycin, were submitted before a CM diagnosis was made [8]. In a second 2019 study, nearly half of 139 enrolled patients diagnosed with CM in metropolitan Phoenix experienced a delayed diagnosis of more than 1 month, and delays were associated with increased healthcare costs [9]. In a third study, Pu et al [10] showed that 73% of 2043 CM diagnoses were made during hospitalization, of which 41% resulted in substantial costs and unnecessary antibacterial administrations. These studies establish the urgent need for the development of early CM diagnostic testing to lessen the number of antibacterial treatments, mitigate the ill effects of broad-spectrum antibiotics, and ultimately improve antibiotic stewardship.

Any rapid point-of-care test offers additional advantages, such as decreased laboratory testing, fewer invasive procedures, reduced patient anxiety, improved outcomes, and lower healthcare costs [8]. Such testing could be made available in remote or underserved locations without immediate access to reference laboratories. The IMMY Sōna coccidioidomycosis LFA assay is the first test designed to meet the need for a rapid CM diagnosis [16]. Although in our study of patients not previously diagnosed with CM the sensitivity was relatively low, the specificity was relatively high. Therefore, further refinement of this rapid LFA test may still be beneficial for patients in clinical settings where positive results might lead to a reduction in or elimination of antibacterials.

Our study is the first to evaluate this LFA assay in a clinical setting for which the test is ultimately intended. It is also the largest prospective coccidioidomycosis study to integrate patients’ serological testing with extensive subanalysis of patient demographics, clinical symptoms, and laboratory values. Our overall rate of CM diagnosis was 17.6%. This is nearly identical to the rate of diagnosis found previously in 1 study that also had a high rate of hospitalized subjects (69%) as did ours (75%) [30]. Although not noted in the previous study, our hospitalized patients received a high number of antibacterial courses prescribed before CM was diagnosed. This has also been reported in another recent study of hospitalized patients with CM [10]. The rate of CM diagnosis in this study is lower than the 29% rate found in another small prospective study [5]. It should be noted this former study only enrolled ambulatory subjects. This raises the possibility that patients with “walking pneumonia” who reside or recently traveled in CM-endemic regions have a higher likelihood of CM responsible for their CAP. In corroborating the previously reported association of low PCT levels in patients with CM [29], our study found that antibacterial therapy might have been avoided in 77% of subjects who subsequently were found to have a positive EIA result if their low PCT had been known. While ESR and CRP were frequently elevated in patients with CM, the merit of blood counts and serum protein levels in CM diagnosis could not be drawn and bears more investigation.

In summary, coccidioidomycosis is a common cause of CAP in the endemic area and among travelers to this region. It presents a diagnostic challenge to physicians and laboratories. A number of diagnostic modalities may be used to distinguish CM from other causes of CAP. Serology is the most frequently used diagnostic test but can be negative early in the course of the disease and does not provide rapid diagnosis in patients. The Sōna coccidioides antibody LFA is an alternative method for a rapid screening approach to early CM diagnosis. In our study we utilized the novel rapid LFA test to improve time-to-result from days to 1 hour. The specificity of this novel LFA test is 92%; however, a sensitivity of 31% demonstrates the need for further investigation and development of an improved screening test. Nonetheless, the development of this LFA and our study’s reinforcement of PCT results to differentiate CM from bacterial CAP are a step forward in earlier CM diagnosis, decreased antibacterial use, and improved antibiotic stewardship.

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 express our appreciation to the patients who participated in this study. We also thank the research coordinators, laboratory technologists, clinical staff, and emergency department personnel at the B-UMCT and B-UMCP for their support.

Financial support. This work was supported by the Centers for Disease Control and Prevention (contract no. 75D30118C02899).

Potential conflicts of interests. J. N. G. is the co-inventor of a recombinant antigen that could be used in a future coccidioidal serologic test. The intellectual property is owned by the University of Arizona and there is currently no product in development with this antigen. 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.