Evaluation of the QuickLab RSV Test, a New Rapid Lateral-Flow Immunoassay for Detection of Respiratory Syncytial Virus Antigen

Abstract

Rapid respiratory syncytial virus (RSV) diagnosis is vital to the prevention of nosocomial RSV infections. We evaluated a new rapid lateral-flow RSV immunoassay, the QuickLab RSV test, that requires use of only one reagent. We compared QuickLab to the Directigen RSV (DIR) assay, which requires six reagents, and direct fluorescent antibody (DFA) testing. DFA results were considered the “gold standard.” For 133 nasopharyngeal aspirates tested, DFA results were 77 (57.8%) positive, 47 (35.3%) negative, and 9 (6.8%) indeterminate. The sensitivities, specificities, positive predictive values, and negative predictive values of QuickLab and DIR tests were 93.3% (70 of 75) and 80.8% (59 of 73), 95.6% (43 of 45) and 100.0% (46 of 46), 97.2% (70 of 72) and 100.0% (59 of 59), and 89.6% (43 of 48) and 76.7% (46 of 60), respectively. QuickLab was significantly (P = 0.02) more sensitive than DIR; the difference in specificities was not significant. DFA was more sensitive than DIR (P < 0.001) but not more sensitive than QuickLab (P = 0.45). The results of DIR testing were initially uninterpretable and required retesting with 15% of the specimens compared to 3% of QL results (P < 0.001). We conclude that the QuickLab RSV test has sensitivity similar to that of the DFA assay and better than that of the DIR assay. QuickLab testing is also simpler to perform and interpret than both DFA and DIR testing.

Rapid diagnosis of respiratory syncytial virus (RSV) infection in infants and children requiring hospital admission can help prevent nosocomial RSV transmission (5, 6, 7), since RSV-infected patients can be either assigned to private rooms or cohorted with other children infected with the same virus. Infection control strategies that include rapid diagnostic tests have been shown to be cost effective for this reason (6). Rapid RSV diagnosis may also be important for therapeutic reasons, although the benefits of ribavirin treatment are unclear (1).

The rapid tests that are now widely used have shown moderately good sensitivity and specificity in studies published primarily in the early 1990s (3, 4, 8, 9, 11). However, these tests generally require multiple processing steps and the addition of reagents. Thus, although total individual test time is short, the hands-on involvement for the operator over the test interval is high. Simpler tests might be amenable to point-of-care use in settings such as emergency rooms. Another potential disadvantage of some existing solid-phase membrane-bound immunoassays is that test results must be read in the same well in which a specimen is absorbed. Substances in the specimen that interfere with absorption may thus make test results difficult to interpret.

The QuickLab RSV (QL) test (Integrated Biotechnology Corp.; test now sold as the Clearview RSV test [Wampole Laboratories]) is a recently approved immunoassay that may overcome these disadvantages. First, it requires the addition of only one reagent to the specimen prior to the addition of a sample to a test strip. Secondly, this test is an example of a newer immunoassay format that relies on lateral flow of antigen-antibody complexes to separate the reading window from the sample well. RSV protein F is detected in the QL test by using a red-colored gold-labeled mouse monoclonal anti-RSV protein F antibody. Protein F antibody complexes travel laterally along the test strip membrane and are detected by a membrane-adsorbed monoclonal anti-RSV F protein at the test line, resulting in a pink- to red-colored line. Unbound or excess mouse anti-RSV protein F passes through the test line and is bound at a control line by a goat anti-mouse immunoglobulin, also resulting in a red line.

Although QL has design advantages over earlier tests, there is little information available about the sensitivity and specificity of the assay. We therefore decided to evaluate the performance characteristics of the QL assay in comparison to those of the Directigen RSV (DIR) test (Becton Dickinson) assay, a membrane enzyme immunoassay that was in use in our laboratory. We used direct fluorescent antibody (DFA) testing and culture as the “gold standard” for assay comparison.

MATERIALS AND METHODS

Nasopharyngeal aspirate (NPA) specimens were collected from pediatric patients at the Children's Hospital of Eastern Ontario, Ottawa, Ontario, Canada, over a 5-week study period in winter of 2003. Following our hospital protocol, a suction catheter of appropriate size was chosen on the basis of patient age. One tube of an aspiration trap was connected to a vacuum source, and the other tube was connected to the suction catheter. The catheter was inserted through the nostril to the posterior nasopharynx, and the aspirate was collected. The suction catheter was then cleared of secretions by suctioning 3 ml of 0.9% sodium chloride into the aspiration trap. Specimens were rapidly transported to the laboratory and kept at 4°C until tested.

Samples were not rejected when blood was present. This was a deviation from the DIR instructions, which state that “excessively bloody” specimens should not be tested. The QL insert does not mandate that bloody specimens be excluded.

Samples obtained during the day and evening were tested with DIR, QL, and DFA at the same time. Samples collected overnight were tested with DIR, refrigerated at 4°C, and tested with QL and DFA the following day. Specimens negative or indeterminate by DFA were set up in culture during day and evening shifts. Testing was performed by lab technicians and technologists in the Eastern Ontario Regional Virology Laboratory.

We performed DIR testing and interpretation according to the manufacturer's instructions. This required the addition of six reagents in total. Initially, 250 μl of a specimen was mixed with the first reagent and then dispensed into the test well and allowed to absorb completely. The other five reagents were added sequentially, and the results were recorded. Total test time was approximately 15 min. DIR results were recorded as positive (purple triangle with white to light-purple background), negative (purple control dot with no triangle), or indeterminate. Several patterns are listed as indeterminate in the product insert: no purple control dot visible; an incomplete purple triangle; a white triangle with a purple surrounding background with a control dot present; and a white triangle with a purple surrounding background without a control dot present. Indeterminate specimens were diluted 1:4 in normal saline and retested.

The QL test was also performed as directed by the manufacturer. This required addition of one extraction reagent to 160 μl of sample, prior to pipetting into the sample well of the test device. Results were read at 15 min. Results were recorded as positive (control and test line seen), negative (control line only), or indeterminate (when the control line had not appeared by 15 min). Specimens with indeterminate QL results were not retested.

DFA was performed by pretreating specimen aliquots with 0.25% N-acetylcysteine (10). Smears were stained with a Merifluor RSV test (Meridian Bioscience Inc.), containing two anti-RSV monoclonal antibodies, according to the manufacturer's instructions. Smears were considered positive when two or more columnar epithelial cells exhibited typical punctate or inclusion-like cytoplasmic fluorescence at high-power magnification (×400). Smears were considered indeterminate when fewer than five columnar epithelial cells were seen at high-power magnification. For all assays, quality control measures were followed per the directions of the manufacturers.

Cell culture was performed by treating specimen aliquots with antibiotics and inoculating onto tube cultures of human fetal lung (derived in this laboratory) and primary rhesus monkey kidney cells. Cultures were incubated for 8 days and observed periodically for RSV cytopathic effect. Isolation of RSV was confirmed using Merifluor reagent. On the basis of an earlier study performed in our laboratory (9), in which DFA detected all culture-positive RSV samples and additionally detected 33% more positive specimens than culture, and other studies indicating that culture is less sensitive than DFA (3), only DFA-negative and -indeterminate specimens were cultured.

For calculation of immunoassay performance characteristics (sensitivity, specificity, positive predictive value, and negative predictive value), samples containing indeterminate results were excluded. Statistical comparisons between assays were done using a McNemar test. This nonparametric statistical test is used to compare paired proportions. For analysis, we considered DFA-positive specimens to be true positives and samples that were DFA negative and culture negative to be true negatives. All reported P values are two-sided and were declared statistically significant when they reached a 0.05 probability level.

RESULTS

A total of 133 NPA specimens were tested. By DFA, 77 specimens were positive (57.8%), 47 were negative (35.3%), and 9 were indeterminate (6.8%). No additional RSV-positive specimens were detected in the cell cultures set up from the DFA-negative or -indeterminate samples. Results of DFA versus QL and of DFA versus DIR are shown in Table 1. The sensitivities of DIR and QL were 80.8% (59 of 73) and 93.3% (70 of 75), respectively. By a McNemar test, sensitivities are statistically different for DIR compared to DFA (P < 0.001) but not for QL compared to DFA (P = 0.45). QL was significantly more sensitive than DIR (P = 0.02). There were nine true-positive samples that were positive by QL and negative by DIR and one true-positive sample that was negative by QL and positive by DIR.

TABLE 1.

DFA compared to QL and DIR RSV immunoassays

Assay and result	No. of samples with DFA result that was:
	+	−	Inda
QL
+	70	2	2
−	5	43	7
Ind	2	2	0
DIR
+	59	0	0
−	14	46	9
Ind	4	1	0

^aInd, indeterminate result.

Specificities were 100.0% (46 of 46) and 95.6% (43 of 45) for DIR and QL tests, respectively. The difference in specificities between DIR and QL was not statistically significant (P = 0.16). The positive predictive values were 100.0% (59 of 59) for DIR and 97.2% (70 of 72) for QL, and the negative predictive values were 76.7% (46 of 60) for DIR and 89.6% (43 of 48) for QL.

DIR results were initially indeterminate with 20 (15%) specimens compared to 4 (3%) of QL results (P < 0.001). With repeat DIR testing after specimen dilution, 15 (11%) were resolved and 5 (4%) remained indeterminate, which was not statistically different from the QL-indeterminate figure (P = 0.72).

DISCUSSION

The new QL lateral flow assay appears to be a robust assay, showing good sensitivity and specificity in the setting of a busy clinical virology laboratory. The higher sensitivity seen with QL compared to DIR will ensure that fewer RSV infections are missed. This will allow physicians to make decisions about initiation of ribavirin therapy in a timely manner. In addition, for hospitals that cohort RSV-infected children, the efficiency of bed allocation will be maximized since fewer patients with false-negative results will be unnecessarily assigned to single rooms. The finding that fewer samples were initially indeterminate with QL than with DIR shows another advantage of the QL assay, since the need for repeat testing of specimens is reduced. Although there were no statistically significant differences in specificity among the three tests, the QL specificity needs to be further studied. A false-positive test with any assay may be clinically important, as it could result in a non-RSV-infected child being cohorted with RSV-infected patients. Reverse transcriptase (RT)-PCR methods could potentially be used to help determine whether any DFA-negative and QL-positive specimens actually contain RSV RNA.

A search of the PubMed database in December 2003 did not identify any published studies regarding the performance characteristics of the QL assay to which we could compare our findings. The DIR sensitivity of 80.8% we found is in line with values reported in the literature, which range from the low of 62% already mentioned to a high of 93%, with various definitions of true positives (3, 4, 8, 11) In a recent study that used two PCR assays to define positivity (4), DIR had a sensitivity of 70.3%. DIR specificity ranged from 86 to 100% in these studies. Our rate of initially indeterminate DIR results (15%) was somewhat higher than the 8.5% reported in an earlier study (11). This may have been due in part to our decision not to exclude bloody specimens, as noted below.

Our study had some limitations. We accepted DFA positives as true positives, although as with all tests there is a potential for false-positive results. There is no perfect gold standard to define RSV positivity, since culture may be insensitive. RT-PCR methods may prove to be the most sensitive technique (4), but not all studies have found high sensitivity. For example, a recent publication found an RT-PCR sensitivity of only 73% in comparison to that of a culture and serology gold standard (2). Falsely negative DFA results are also a possibility.

As well, our study protocol deviated from assay insert instructions in two instances. We did not exclude bloody specimens from DIR testing, although assay insert instructions recommend these be excluded since false positives or indeterminate results may be obtained. Inclusion of these samples may have led to an increase in the proportion of indeterminate results with DIR, but we elected to do so to increase the generalizability of the study to real-world conditions and so that all specimens could be compared with all three tests. Unfortunately, we were not able to quantify accurately the number of “excessively bloody” (as worded in the DIR insert) specimens tested. A qualitative written description of each specimen was recorded, but descriptions from individual technologists of the amount of blood present were later found to differ widely. A quantitative measure or score would have allowed us to analyze our results with regard to the effects of blood contamination; that approach is recommended to others planning RSV assay comparisons.

We also elected not to repeat tests of QL-indeterminate or DFA-indeterminate specimens. Our present policy is to issue a recommendation for repeat sample submission for indeterminate DFA samples rather than retest the original sample, since we feel that an indeterminate result would likely again be obtained. We also made this assumption with the QL test. However, it is possible that a valid result could have been obtained by repeat testing of a different portion of the original specimen with DFA or QL. Future studies of the QL assay should assess whether a strategy of retesting indeterminate specimens is diagnostically and economically preferable to one of requesting a new sample after a single indeterminate result.

Finally, although we anecdotally noted that hands-on time was shorter with the QL test than with DIR, since only one reagent is added rather than six, we did not attempt to quantify the time difference.

In summary, although the DIR assay has high specificity, we feel that its poorer sensitivity and its relatively high frequency of indeterminate results that require specimen retesting make the QL assay a better choice as a rapid test.

The DFA test remains an excellent method for RSV diagnosis but takes much longer to perform than rapid immunoassays, and a highly skilled technologist is needed to interpret the results. QL may thus be a good option for laboratories lacking DFA expertise or where DFA testing is not available at all times. Given its rapidity and ease of use, QL may also be a useful point-of-care test in the emergency room or hospital ward setting, although this must be studied prospectively. Future studies are also required for evaluation of the QL test in comparison to other newer diagnostic tests for RSV, such as RT-PCR and other recently developed lateral flow immunoassays.