Sensitivity of nasopharyngeal, oropharyngeal, and nasal wash specimens for SARS-CoV-2 detection in the setting of sampling device shortage

Eligible participants were ≥ 18 years old and hospitalized in the internal medicine wards at the Geneva University Hospitals, who had a positive SARS-CoV-2 rRT-PCR in a NPS specimen in the preceding 1–6 days. ICU patients were excluded.

PCR tubes (CobasTM Roche Reference No. 07976577001-3) were filled with 3 ml of DMEM. Nylon flocked NPS (COPAN Reference No.A305CS01) and cotton OPS (VWR Reference No. 300260) were used for sampling. Consecutive OP and then NP samples from 29 patients were obtained in parallel and inserted into DMEM Cobas tubes. The OPS was always performed first. OPS specimens were obtained by swabbing the oropharyngeal posterior wall and turning once in each direction. They were then transferred into the Cobas PCR tube. NPS were performed according to the usual technique [4, 11]. Specimens were stored at 4 °C after being collected.

PCR tubes (CobasTM Roche Reference No. 07976577001-3) were filled with 1 ml of DMEM and 2 ml of NaCl 0.9% was added in half of them. Consecutive NW and NPS specimens from 20 volunteers were obtained. NW was always performed first, and as follows: 3 ml of sterile saline solution was injected into the nostril using a 3-ml syringe and recovered into a plastic cup by leaning patients’ heads forward. Using the same syringe, a total 2 ml volume of the NW was transferred into a Cobas PCR tube containing 1 ml of DMEM media. NPS were performed according to the usual technique [4, 11]. They were then transferred into a Cobas PCR tube containing 1 ml of DMEM media and 2 ml of NaCl 0.9% in order to compare the two techniques using equal media dilutions. Specimens were stored at 4 °C after being collected. Video demonstrating the NW procedure is available at https://​www.​youtube.​com/​watch?​v=​3cMoR7hSPF8&​feature=​emb_​title.

Viral RNA genome detection was performed by RT-PCR using the Roche Cobas 6800 system (Cobas SARS-CoV-2 Ref 09175431190; Cobas SARS-CoV-2 Control kit Ref 09175440190; Cobas 6800/8800 Buffer Negative Control kit Ref 07002238190). This technology allows nucleic acid extraction, purification, PCR amplification, and detection of SARS-CoV-2, targeting ORF1a/b and a pan-sarbecovirus conserved region of the E-protein gene.

We assessed the suitability of the Dulbecco’s modified Eagle medium (DMEM) for use in testing specimens by comparing it with the universal transport medium (UTM). First, ten positive NPS specimens were simultaneously spiked in 3 ml DMEM and 3 ml UTM and then analyzed by rRT-PCR using the Roche Cobas 6800 system. Then, to further evaluate the sensitivity of the UTM, a patient positive sample showing a Ct value around 33 was serial diluted in DMEM or UTM transport media and then analyzed by rRT-PCR using the Roche Cobas 6800 system.

The correlation between the Ct values for ORF1 (arbitrary chosen for comparison) in NPS and in OPS or in NW specimens was evaluated using the Pearson correlation coefficient (r). The negative specimen by NW was arbitrarily assigned a Ct value of 45 and the two negative specimens by both OPS and NPS were excluded from the analysis. Correlation was also represented graphically using a simple linear regression (Figs. 1 and 2). Statistical analyses were performed using IBM SPSS Statistics 25.

The resulting cycle threshold (Ct) values were very similar when comparing the DMEM with the UTM. The delta Ct values ranged from 0.01 to 2.59 with a mean delta Ct value of 0.53 for ORF-1 and 0.67 for E-gene (supplementary data, table 1A). Both media seem equivalent for detection of low SARS-CoV2 loads (supplementary data, table 1B).

We compared the techniques in two groups of volunteers: the first group comprised 20 cases where a NW specimen was collected followed by a NPS specimen. The second group comprised 29 cases where an OPS specimen was collected followed by a NPS specimen. The clinical sensitivities of NW and OPS specimens were compared with those of the NPS specimens using the Ct values obtained by SARS-CoV-2 rRT-PCR (supplementary data, table 2). Out of 20 cases, one patient that was positive with the NPS sampling with Ct values of 33.46 for ORF1 and 35.12 for the E-protein gene had negative results with the NW sampling. When comparing the NW sampling with NPS sampling, the mean delta Ct values were 1.77 (range − 6.82–7.06) and 1.73 (range − 7.79–8.25) for ORF1 and E-protein gene respectively (supplementary data, table 2). The Pearson r was 0.75 (p < 0.01), showing a statistically significant correlation between the ORF1 Ct values for the NPS and the NW specimens (Fig. 1).

Out of 29 patients, two cases were negatives in both the OPS and the NPS specimens. When comparing OPS with NPS sampling, the mean delta Ct values were 1.24 (range − 4.24–5.8) and 1.32 (range − 4.63–7.6) for ORF1 and E-protein gene respectively (supplementary data, table 2). The Pearson r was 0.88 (p < 0.01), showing a statistically significant correlation between the ORF1 Ct values for the NPS and the OPS specimens (Fig. 2).