A narrative review of nine commercial point of care influenza tests: an overview of methods, benefits, and drawbacks to rapid influenza diagnostic testing

Methods

Results

Sensitivities and specificities reported for these nine commercially available RIDTs were determined from the manufacturers’ conducted studies, independent peer-reviewed trials, and meta-analyses published within the last 10 years (Table 1). Table 2 describes the names, specimen types, sensitivity, and specificity for these tests against both IAV and IBV detection [23, 34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46]. Whereas manufacturers reported their sensitivity and specificity to be on the higher end of the ranges, the meta-analyses and other peer-reviewed trials reported many studies having lower sensitivities and specificities, producing the ranges of values for these categories denoted in Table 2 [29, 33].

When conducting a PubMed search for any publications calling out the commercial RIDTs by name in either the abstract or the title, the Veritor, Xpert Xpress, BinaxNOW, and Cobas NAAT tests had significantly more results than the remaining five tests. Abbott’s BinaxNOW Influenza A and B has the most named studies in the peer-reviewed literature, with 37 results appearing on PubMed following our search criteria, while Status Flu A&B had the fewest, with no peer-reviewed references under our inclusion criteria (Table 1).

Table 3 demonstrates the products along with their type of testing method and the time to results reported by the manufacturers [32, 33, 41], [42], [43], [44], [45], [46]. The table shows that one of the newer RIDTs, known as ID NOW Influenza A & B, can obtain results in the shortest period of time, 5–13 min, whereas the RT-PCR gold-standard method produces results in roughly 30 min [3]. On average, the most commonly utilized methods in the clinics today are LFAs that take 10–15 min on average to produce results (Table 3) [2, 5], [6], [7, 24].

Discussion

Herein we compiled a thorough list of the commonly utilized, commercially available RIDTs from the US CDC’s listing of FDA-approved influenza diagnostic tests to provide an overview of the molecular methods, sensitivities, specificities, and time to results. As seen by the high variability in test sensitivity in Table 2, the many studies incorporated in this review have shown that the reported sensitivities and specificities by manufacturers’ protocols are not always comparable to what is observed in a variety of clinical and public health settings. For every assay examined in Table 2, the upper end of the sensitivity and specificity ranges reported comes from the manufacturers’ own reported data on the tests, with all nine reporting IAV and IBV sensitivity and specificity greater than 90.0% when compared to RT-PCR or viral culture [32, 33, 41], [42], [43], [44], [45], [46]. However, as seen in Table 2, the range of reported sensitivities stretches far below the manufacturers-reported 90.0% for every test in both IAV and IBV detection. This implies that the real-world data obtained from the peer-reviewed studies included in this review demonstrates that the clinical performance of these assays does not always mirror the initially reported data by the manufacturers. The large disconnect between the data reported from the manufacturers on the clinical performance of their assay and the data that are reported in the peer-reviewed literature is not to be overlooked. This significant overestimation of performance by the manufacturers could result an in a high rate of false negatives that directly impact patient and community health.

However, there is a critical need for expanded utilization of RIDTs in both the public health and primary care fields, with accurate viral diagnostics becoming the cornerstone for infection prevention and quick and appropriate therapeutic initiation [10, 15]. Although sensitivity and specificity are shown to greatly fluctuate with many of the RIDTs, the utility of rapid testing with the potential for false readouts still outweighs foregoing testing or waiting days for results to come in, especially when dealing with a viral illness in which time is critical for appropriate antiviral initiation [15]. It is our opinion that these technologies, while having technical limitations, provide a strong benefit to increasing accurate viral diagnostic testing capabilities. The thermostability and the low cost associated with many of these tests makes them ideal candidates for field implementation or utility in primary care offices, yet neither of these settings may have access to the advanced pathology testing laboratories that typically house the large-scale PCR equipment utilized in viral diagnostics.

Of the influenza testing methods discussed, conventional RT-PCR remains the gold standard due to its excellent sensitivity and specificity [3]. Although this method of accurate detection of influenza is clinically important, there are many difficulties to overcome in implementing traditional RT-PCR workflows into medically rural and underserved communities where primary care and public health efforts are often at work attempting to provide accurate viral diagnostic strategies. Due to the cumbersome equipment and supplies associated with RT-PCR, the financial burden on both the clinical practice and patient can quickly mount when having to either establish a CLIA testing lab for in-house sample processing or specimen transport to an external testing lab. This has led to a hybrid type of method in which many primary care clinics perform their testing in-house with the LFAs or other RIDTs. However, due to the known variability in sensitivity of the current commercially available RIDTs, negative test results obtained via RIDTs for a patient with a clinical presentation highly suspicious for influenza are often sent for confirmatory testing via conventional RT-PCR or viral culture [2, 6]. While this is a more economical solution than having to house the equipment for RT-PCR, the time to receive a confirmatory test is still slower due to transit time and the time to perform the test, ultimately delaying the appropriate treatment and care for the patient [7, 9]. However, if the test is positive, studies have shown that due to the high specificity of LFAs, the likelihood of being infected with the virus is extremely high [37], [38], [39]. Additionally, LFAs are roughly $15-$20 per test, which makes it financially reasonable to perform a quick test [37]. Through investigation via the study’s PubMed inclusion criteria, individual studies document that LFA sensitivities for IAV are variable, ranging from 4.4 to 100% (Table 2) [32, 33, 41], [42], [43], [44], [45], [46]. Although this is a wide range, a meta-analysis from 2017 evaluating 162 publications showed that traditional RIDTs reported a pooled sensitivity and specificity of 54.4% (95% CI, 48.9–59.8%) and 99.4% (95% CI, 98.9–99.8%), respectively for IAV, with 53.2% (95% CI, 41.7–64.4%) and 99.8% (95% CI, 99.7–99.9%), respectively, for IBV [37]. Among the RIDTs currently on the market, most are equipped with a digital reader that allows for more accurate readings when the test lines may be too faint to provide a reliable human interpretation of results [33, 35, 37]. Although the specificities are high, the wide range of sensitivities can lead to many errors in the clinical setting, such as the inappropriate use of antibiotics, the underprescribing of antivirals, the increased cost for confirmatory testing, and increased hospitalizations. In order to combat the high rates of misdiagnosed influenza cases, the FDA placed a minimum sensitivity range of 80.0% compared to RT-PCR on RIDTs in 2017 [47].

Another methodology gaining popularity in the RIDT market are nucleic acid amplification tests (NAATs), such as the Abbott ID Now platform that reported sensitivities and specificities greater than 90.0% in the reported data (Table 2). The utility of this technology is growing in its application given the accuracy of results similar to that of RT-PCR while reducing the time and resources needed to achieve viral detection [29, 33, 48]. As a NAAT-based assay, ID NOW IAV and IBV can run a test in as few as 5 min with a negative read at 13 min [32]. The quick turnaround time from this rapid-detection technology in clinical settings results in one-third the original time could drastically decrease the exposure time to other patients or staff, providing a theoretical reduction in community propagation secondary to exposure. However, the increased detection capabilities associated with NAATs also demonstrate an increased sensitivity to contamination. It has been documented that NAATs carry a high risk of false positives due to cross-contamination in the machine or preparation space with persistent genetic material or amplicons from previous samples [49]. It is essential that the spaces utilized for conducting these tests are thoroughly cleaned between samples in an attempt to minimize this risk of false positives for point-of-care testing.

In addition to the logistical and financial considerations discussed previously, our review of the literature has highlighted other factors worth noting that may impact the reported sensitivity and specificity of the RIDTs; antigenic variation, sample population demographics, and clinical sample type are all factors to consider in point-of-care influenza testing. Clinicians, researchers, and public health officials should all consider these factors when determining if an RIDT would appropriately fulfill their needs.

Conclusions

As seen in the current global pandemic caused by COVID-19, the need for accurate and accessible viral diagnostic tests is great [53]. Herein, we have conducted a review of nine of the commonly utilized commercial RIDTs. The molecular basis of each test, including the assay target and the time to diagnosis, was shown as reported by the manufacturer (Table 3). Additionally, the sensitivity and specificity of each assay has been reported in Table 2.

Through this review of the literature, we recorded the fastest diagnostic modality of the RIDTs to be the NEAR technology utilized by the Abbott ID Now Influenza A&B test, recording positive results in as little as 5 min, with the longest time being the RT-PCR–based Cepheid Xpert Xpress Influenza A/B taking up to 30 min to finalize a negative result (Table 3). Of the nine RIDTs examined in this review, six of the nine reports finalized a result within 15 min or less (Table 3). The lateral flow assays and NEAR tests report faster results when compared to the RT-PCR based tests; however, the inverse seems to be the case with regard to reported sensitivity and specificity for detection of IAV and IBV. The RT-PCR–based assays outperformed the LFA and NEAR RIDTs, averaging sensitivity and specificity above 90.0% for all reported data (Table 2).

At the time of this submission, we did not identify any recent review articles compiling this information, including the manufacturer’s reported data on each of the RIDTs. It is our hope that this narrative review on the current RIDT technology can be utilized by clinicians and public health professionals as a resource for determining the appropriateness of a given assay for their patients. While we acknowledge that more real-world research is needed around these assays and their diagnostic accuracy for various IAV and IBV strains, this compilation of the current research surrounding these nine commercially available RIDTs should serve as a robust source of information on the technology employed by each test as well as the current reported statistics on their performance.