A Rapid Test to Differentiate Viral From Bacterial Infections: Searching for the Holy Grail

TL;DR

This paper reviews a rapid test combining two biomarkers to help distinguish viral from bacterial infections and reduce unnecessary antibiotic use.

Contribution

The paper evaluates a combined point-of-care assay using Myxovirus resistance protein A and C-reactive protein for infection differentiation.

Findings

The combined assay shows potential for rapid differentiation of viral and bacterial infections.

Remaining questions include the test's accuracy and applicability in diverse clinical settings.

Abstract

Differentiating viral from bacterial causes of acute respiratory illness can avoid the numerous consequences of inappropriate antibiotic use. We review the clinical literature evaluating a combined rapid Myxovirus resistance protein A and C-reactive protein point-of-care assay, discuss potential applications, and highlight remaining questions.

Tables

Table 1.. Studies Evaluating the Performance of a Combined Point-of-Care Myxovirus Resistance Protein A and C-Reactive Protein Assay for Detecting Bacterial Infections

Year	Author	Country	No. of Participants	Clinical Presentation	Age in Years (Mean)	Reference Testing Method	Predicting Bacterial Infectionc		Ref.
							Negative Predictive Value (95% CI)	Positive Predictive Value (95% CI)
2015	Sambursky	United States	54	Febrile pharyngitis	18–69 (37)	PCR, bacterial culture, urine antigen, chest X-ray	80% (56%–94%)	88% (73%–97%)	[24]
				Febrile lower respiratory symptoms	22–29 (51)
2017	Self	United States	205	Febrile upper respiratory tract symptoms	Median 29 (IQR 31)	PCR, antibody serology, bacterial culture, procalcitonin, white blood cell count	97% (94%–99%)	63% (45%–79%)	[25]
2018	Shapiro	United States	220	Febrile upper respiratory tract symptoms	2–86 (37)	PCR, antibody serology, bacterial culture, procalcitonin, white blood cell count	99% (93%–100%)	76% (59%–87%)	[26]
2022	Shapiro	United States	520	Recently (within 3 d) febrile with acute respiratory symptomsa	1 to >65 (35)	PCR, antibody serology, bacterial culture, procalcitonin, white blood cell count	99% (97%–99%)	58% (49%–67%)	[27]
2023	Tong-Minh	Netherlands	224	Individuals with respiratory symptoms or who had throat swabs collected	Median 58 (IQR 28)	PCR, bacterial culture, procalcitonin, chest imagingb	89% (82%–94%)	60% (51%–68%)	[23]

Full text

Acute respiratory infections led to over 13 million emergency department visits in the United States in 2022 [1]. More than 25% of such patients with bronchiolitis between 2007 and 2015 were treated with antibiotics, of which 70% had no documented bacterial infection [2]. Of antibiotics prescribed to privately insured patients with acute respiratory illnesses in 2016, 23% were inappropriate [3]. Inappropriate antibiotic use among children in the United States was associated with an increase in 7-day adverse events and an estimated $24.5 million in attributable expenditures (2025 US dollar) [4]. Exposure to antibiotics early in life may increase the risk for obesity and asthma [5, 6] as well as drive antimicrobial resistance—a major global public health threat [7]. Thus, effective strategies to assure the appropriate use of antimicrobials among cases of acute respiratory illness are urgently needed.

For decades, researchers have sought a means to distinguish bacterial from viral infections. C-reactive protein (CRP) has been used since its discovery in the 1930s. In the early 1990s, procalcitonin was introduced; however, while more accurate than CRP alone [8], procalcitonin was shown to be insufficiently sensitive for some use cases [9]. More recently, a score using CRP and 2 other cytokines demonstrated promising discrimination [10]. Yet, such tests are not widely available. We describe the biologic mechanisms of a point-of-care lateral flow assay combining CRP and Myxovirus resistance protein A (MxA) to distinguish bacterial from viral infections, highlight available clinical performance data, discuss potential applications, and consider areas for further research.

C-REACTIVE PROTEIN AND MYXOVIRUS RESISTANCE PROTEIN A

CRP was discovered in the 1930s among patients with pneumococcal pneumonia. C-reactive protein is an acute phase reactant produced by the liver and stimulated predominantly by mediators of the innate immune system, specifically interleukin 6 [11]. C-reactive protein rises within 4–6 h of inflammation and peaks within 36 h [12]. C-reactive protein has been studied as a biomarker to identify numerous infectious states, with higher values (above 80 mg/L) thought to represent more severe infections [12]; however, CRP may be less accurate than procalcitonin for distinguishing bacterial from viral infections [8]. Yet, CRP is readily measured, making it an appealing biomarker for point-of-care assays.

Myxovirus proteins were discovered in the 1960s among mice with resistance to influenza viruses [13]. Related genes were later discovered in humans encoding MxA [14]. Myxovirus resistance protein A is expressed in response to type 1 interferon [15]—a mediator of the innate immune system's specific response to viral infections [16], which does not rise in response to even severe bacterial infections [17]. Myxovirus resistance protein A is exclusively stimulated by interferon type 1 and not by pathogens or pathways of response to bacterial infection (eg, interleukin-1 or tumor necrosis factor alpha) [18]. In vitro studies have shown that levels of MxA rise within 1–2 h of interferon induction and have a half-life of approximately 2 days [19]. Mechanistically, MxA inhibits transcription and replication of a broad range of viruses [18].

Numerous studies have evaluated MxA as a marker of viral infection [20]. MxA levels were significantly higher among 95 children with viral infections determined by either polymerase chain reaction (PCR) testing or clinical judgement compared to 27 children with bacterial infections or 52 age-matched uninfected controls [21]. Among 247 children with symptoms of lower respiratory tract infection, 5 with bacterial infections, MxA detection had an area under the receiver operating characteristics curve of 90% for classifying viral versus bacterial etiology [22]. In a prospective cohort study of 193 children, MxA levels demonstrated a 96% sensitivity for distinguishing viral infections from bacterial infections and uninfected controls [23]. Those and other studies demonstrated that elevated MxA was associated with infection with RNA viruses such as influenza A and B, parainfluenza virus, respiratory syncytial virus, coronavirus, including SARS-CoV-2, and human metapneumovirus [23, 24], as well as with DNA viruses such as Epstein–Barr virus, herpes simplex virus types 1 and 2, cytomegalovirus, and adenovirus [24].

However, while a positive MxA test might indicate the presence of a viral infection, neither a positive nor a negative result excludes a bacterial infection. Simultaneous detection of a CRP value below 20 mg/L would likely indicate a viral infection, whereas a negative MxA result and detection of CRP above 20 mg/L [12] could indicate a non-viral illness—which in the right clinical context could warrant antibiotic therapy. The FebriDx® assay (Lumos Diagnostics, United States) is a Food and Drug Administration–cleared test that incorporates qualitative detection of both MxA (≥40 ng/mL) and CRP (≥20 mg/L) into a disposable lateral flow device and provides results after 10 min.

CLINICAL PERFORMANCE OF A COMBINED MYXOVIRUS RESISTANCE PROTEIN A/C-REACTIVE PROTEIN TEST

The MxA/CRP assay has been evaluated across numerous contexts (Table 1). A prospective single-center trial among 54 patients with acute, febrile respiratory tract illness found that the MxA/CRP assay correctly identified 80% of individuals with a bacterial infection confirmed with PCR, positive cultures, bacterial-specific urine antigens, or elevated bacterial-specific IgM antibodies [24]. Negative MxA/CRP results also correctly identified 92% of individuals with no infectious cause of their symptoms [24].

Another prospective multicenter study among 205 patients compared the MxA/CRP test to a reference testing algorithm that included throat culture, PCR testing from an upper respiratory tract sample, procalcitonin, white blood cell count, bandemia, and clinical judgement. That study found a 92% agreement for bacterial infection (sensitivity 80%, specificity 93%) [25]. The MxA/CRP assay demonstrated a 92% overall agreement with reference testing and/or physician panel adjudication in a subsequent prospective multicenter trial among patients with febrile upper respiratory tract infections [26]. Similar results were reported by a multicenter study of 520 patients presenting with febrile acute respiratory tract infections to outpatient centers; the MxA/CRP assay had a 93% sensitivity and 88% specificity for bacterial infections compared to a reference testing algorithm that included PCR and culture for 28 different pathogens as well as serum evaluation of procalcitonin, bacterial and viral antigens and white blood cell counts [27]. Finally, in a real-world study among adults, the MxA/CRP assay had an 87% sensitivity and 64% specificity for bacterial infection; of note, 93% of participants in that study had at least one comorbidity, nearly 30% were using immunosuppressive medications, and 30% were asymptomatic [28].

The MxA/CRP results also impact management. A single-center retrospective study in the primary care setting found that MxA/CRP testing led to deferral of antibiotics in 8 of 12 (67%) patients for whom bacterial respiratory tract infections were initially suspected [29]. Among 216 children with symptoms of an acute respiratory illness, MxA/CRP results led to a change in the therapeutic plan and a decision to defer antibiotics in 10% of cases [30]. Among 162 participants from 9 outpatient practices with symptoms of a lower respiratory tract infection deemed likely to receive antibiotics, MxA/CRP results led to a 41% reduction in antibiotic use with no increase in representation to care over 28 days of follow-up [31]. A cost analysis estimated that an MxA/CRP assay with a 95% sensitivity for bacterial infections would save over 2.5 billion dollars per year in the United States alone, with savings persisting even with a sensitivity of 75% [32].

The MxA/CRP test has also been used to triage patients. Among 1321 patients with symptoms of COVID-19, 84% were negative by MxA/CRP, of which 78% were triaged to a low-risk area leading to a significant reduction in use of emergency department resources [33]. Elsewhere, MxA/CRP results permitted 826 patients with suspected COVID-19 to be released from quarantine who would have required quarantine based on clinical criteria [34]. Notably, in that study MxA/CRP results missed 10 of 136 cases of SARS-CoV-2 infection detected by PCR, which were either mildly symptomatic or asymptomatic.

REMAINING QUESTIONS AND FUTURE DIRECTIONS

While available performance data are promising, additional questions remain. How MxA/CRP testing performs in other settings, such as intensive care units or inpatient wards, as well as for less common viruses is not known. Similarly, viral evolution might alter the immune response. One study found that the MxA/CRP assay had a notably lower sensitivity for viral infections (56%) during a period when a novel strain of SARS-CoV-2 was the predominant circulating variant [35].

Further, no gold standard diagnostic exists that can distinguish bacterial from viral infections against which to compare the MxA/CRP results. Molecular assays have high sensitivity for specific pathogens; however, currently available diagnostics are not sufficiently robust to definitively rule out all viral or bacterial causes of respiratory illnesses. Thus, individual study-level performance data should be interpreted within the context of the existing diagnostic limitations.

Additionally, more than 3% of adults and up to 15% of children with bacterial pneumonia may have a detectable viral pathogen [36, 37]. In such instances, positive MxA and CRP results may lead to inappropriate discontinuation or deferral of antibiotics in the absence of other clinical manifestations of a bacterial infection. Similarly, some viral infections have been shown to raise CRP values above 100 mg/L [38], which will likely still result in inappropriate antibiotic use among a minority of cases. Furthermore, CRP peaks after 36 h, and thus false-negative values may occur early in the disease course. For those reasons, clinical context and follow-up remain essential; however, those considerations do not necessarily negate the impact of rapid MxA/CRP testing. Further research is needed to delineate the limitations of the test. Additionally, more evidence is needed on clinical outcomes for patients treated based on results of the MxA/CRP assay.

Thus far, the MxA/CRP assay has been utilized predominantly in emergency department and primary care settings. Additional areas that may benefit from such tests include nursing homes and long-term care facilities. More than half of residents in some long-term care settings may be prescribed antibiotics within a 12 month period [39]. Similarly, antibiotics are prescribed in nearly 40% of urgent care visits and nearly half of inappropriate antibiotic prescriptions occur in urgent care centers [40]. Thus, substantial benefits from antimicrobial stewardship may be appreciated if rapid MxA/CRP testing is introduced in such settings. However, further research is needed to confirm that the test performs as expected in other settings and among other populations.

CONCLUSIONS

Combined MxA/CRP testing can reliably exclude bacterial infections in acute respiratory illness. The high negative predictive value provides confidence for clinical decision making. However, important questions remain, including delineating longitudinal clinical outcomes. Along with other rapid point-of-care tests to detect specific pathogens, the MxA/CRP test may finally provide a way for frontline clinicians to objectively determine the need for antimicrobials among cases of acute respiratory illnesses, the importance of which cannot be overstated.