Diagnostic accuracy of VIKIA® Rota-Adeno and Premier™ Rotaclone® tests for the detection of rotavirus in Niger

Rotavirus gastroenteritis represents an important public health concern, especially in low-income countries with an estimated 215,000 annual deaths among children under 5 years of age primarily in Sub-Saharan Africa [1]. The WHO recommends the use of enzyme immunoassays (EIA) carried out in hospital laboratory facilities for rotavirus diarrhea surveillance [2]. In areas with decentralized health systems, hospital surveillance might miss an important proportion of diarrhea cases, which are often managed at the level of primary health centers [3, 4]. Initial evaluations of rotavirus detection by EIA and RDT mainly used culture, electron microscopy or viral RNA electrophoresis as the gold standard with good sensitivity and specificity [5, 6]. Recent evaluations using highly sensitive reference methods as reverse transcription polymerase chain reaction (RT-PCR) have shown lower sensitivity of RDTs and EIAs for the detection of group A rotavirus in stool specimens [7, 8]. Few studies have evaluated RDT and EIA tests in parallel to compare their performance on the same samples and showed similar performance for both tests [8, 9]. While some studies reported a high level of false positive results with the widely used RDT VIKIA® Rota-Adeno (bioMérieux, Marcy l’Etoile, France) [8, 10], other studies showed good sensitivity and specificity [7, 11].

We conducted surveillance of rotavirus in hospitals and health centers in Niger from 2010 to 2012 using the RDT VIKIA Rota-Adeno assay, which was selected based on the good performance data published at the time, and the possibility to use the test in decentralized settings. Considering the concerns raised about this test afterwards, we subsequently performed a parallel evaluation of the diagnostic accuracy of VIKIA® Rota-Adeno RDT and one of the WHO recommended EIA, the Premier™ Rotaclone® (Meridian Bioscience Inc, Cincinnati, Ohio, USA), compared to RT-PCR as reference standard, using specimens collected in the surveillance project.

The performance of VIKIA® Rota-Adeno and Premier ™ Rotaclone® tests using RT-PCR as gold standard illustrated in Table 1 showed equal sensitivity of 80.7% for both assays and a respective specificity of 95.5 and 100% for the RDT and EIA tests.

The sensitivity of both assays was in accordance with the results of previous studies using RT-PCR as reference standard [7–9]. This reduced sensitivity compared to what has been reported in the manufacturer inserts for both assays could be due to the lack of detection of certain genotypes by EIA and RDT. However, the 23 false negative specimens by both EIA and RDT were genotyped and showed the most frequent genotypes in our study (10 G2P[4], 8 G12P[8], 2 G3P[6], 2 G3+G9P[6] and 1 G9P[8]) [3], suggesting no clear association between lack of detection and particular genotypes. Instead, it is most probably due to the high sensitivity of the reference standard [8, 9] and notably to the ability of RT-PCR to detect samples with low viral loads that cannot be detected by RDT and EIA assays [7, 12]. Unfortunately, the rotavirus viral load in samples was not assessed during this study to confirm this hypothesis. However, the relationship between viral load (assessed by Cq value) and EIA positivity has been well established by other groups [12, 13]. Low viral loads of rotavirus were also found in healthy controls, which led these authors to conclude that RT-PCR-positive EIA-negative samples do not reflect illness attributable to rotavirus and that EIA is a better suited method for laboratory diagnosis of rotavirus-associated diarrhea.

While specificity of Premier™ Rotaclone® was perfect in our study, the specificity of VIKIA® Rota-Adeno was lower, but still acceptable for surveillance purposes. The specificity of VIKIA® Rota-Adeno found in our study was much better than the specificity reported in an Australian study (54.3%) [8]. The low specificity found there might be due to a momentary low quality of the product since other evaluations either before [9, 14], or after, using tests produced after optimization of the manufacturing process [7], reported a good specificity for this test. In our study, all six VIKIA® Rota-Adeno false positive results showed a weak signal, compared to 15 of 96 (16%) weak positive results among positive samples by RT-PCR. As with other RDTs, although weak lines could represent false positive results, any (colored) line should be considered positive for optimal sensitivity of the rapid test. Strict respect of the time for results reading of RDT could help to overcome misreading, which is 10 min in the case of VIKIA® Rota-Adeno test.

Thus, considering the similar sensitivity of the VIKIA® Rota-Adeno and Premier™ Rotaclone® compared to RT-PCR in this study and the lower but acceptable specificity of the RDT, the rapid test could be a good alternative for use in peripheral health centres where laboratory capacity is limited.

The main limitation of this study was the use of thawed specimens. Frozen samples might slightly change the test performance, as illustrated by the fact that 6 of the 23 false negative specimens in this sub-study were initially positive by RDT in the surveillance activities, while one initially false-negative sample was found positive when testing frozen specimens. However, the correlation between the results of the VIKIA® Rota-Adeno test in the field on fresh samples and results on frozen samples was good (kappa = 0.88), similar to the correlation between field and central laboratory results for quality control during the surveillance study (kappa = 0.83), suggesting that testing on frozen samples did not strongly affect test results.