gms | German Medical Science

The severe acute respiratory syndrome (SARS) emerged in southern China in November 2002 and spread in an epidemic manner. We have identified a novel coronavirus (SARS-CoV) as the causative agent by a combination of virus culture, low stringency random RT-PCR, and immunofluorescence assay. The virus grew on Vero cells with a clear cytopathic effect. Immunofluorescence showed seroconversion in an index patient after defeverescence. Random RT-PCR yielded two different amplified fragments of 90 and 300 nucleotides, respectively, from the cell culture supernatant. Phylogenetic analysis rendered the new virus to be only distantly related to other coronaviruses.

Several molecular tests were established upon identification of SARS-CoV, one of which has been transformed into the first commercial diagnostic kit for SARS-CoV (realArt HPA coronavirus detection kit, Artus-Biotech). The kit was as sensitive as nested RT-PCR, as determined in examination of 161 samples of confirmed and suspected SARS patients from Hanoi, Vietnam. PCR inhibition occurred in a substancial fraction of sputum and stool samples (30 and 25%, respectively), and nasophanryngeal swab samples contained significantly less RNA than the aforementioned. Statistical analysis revealed that even the highly optimized kit failed to detect virus in all samples that were actually virus-positive.

To increase the sensitivity, a real-time RT-PCR assay for nucleocapsid gene subgenomic RNA was established, which is highly abundant in cells. In a cohort of 66 specimens from German, Singaporean, and Hong Kong patients with confirmed SARS, the N assay was compared against commercial real-time RT-PCR kits from Roche and Artus. Sensitivities of all tests were not significantly different and ranged around 70% in the total of samples.

Culture cells freshly infected with SARS-CoV contained five times more subgenomic N RNA than genomic RNA, but the respective levels converged during four days, explaining why in clinical samples the predicted excess of N RNA was absent (measured by real-time RT-PCR).

RNA concentrations in different sampling compartments decreased in the order sputum/trachea ≥ stool ≥ throat swabs, and all (n=12) samples from the lower respiratory tract tested positive in all tests.

In conclusion, the widely used swab and stool samples have to be considered inappropriate material for SARS-CoV detection. Provoked sputum is highly favorable and might enable ruling out SARS by RT-PCR, but due to the infection risk entailed by sampling its application will remain restricted to high security isolation wards.