nasopharyngeal samples were collected in plastic tubes containing 2mL of ultrapure water. A total of 388 nasopharyngeal samples were included in the study ( Supplementary Table 1 ). Those samples were randomly selected as the SARS-CoV-2 positivity rate was really high during the period this study was done. Those samples followed two sequential SARS-CoV-2 diagnosis protocols as illustrated in Figure 1 : a) the standard RNA extraction using commercial column kits; b) a mechanical lysis using a homogenizer without RNA extraction.

Homogenization was performed in the Bead Ruptor Elite equipment (OMINI International, USA) whose operation is based on the vigorously shaking of the vials that contains sample at a set speed and time to lyse cells. Despite being a bead mill homogenizer, no beads are used in this processing. The homogenization program is regulated at 4 m/s for 30 seconds.

RNA extraction using the Virus DNA/RNA Purification kit (Biocomma, China). RNA extraction was carried out using this commercial kit by following the manufacturer’s manual.

Both the homogenized samples and the RNA extractions were processed for SARS-CoV-2 detection by RT-qPCR as we have previously reported (Freire-Paspuel and Garcia-Bereguiain, 2021a; Freire-Paspuel and Garcia-Bereguiain, 2021b; Freire-Paspuel and Garcia-Bereguiain, 2021c; Freire-Paspuel et al., 2021; Santander-Gordon et al., 2021; Freire-Paspuel et al., 2022; Rodriguez-Paredes et al., 2022). The RT-qPCR was performed in a the CFX96 (Biorad, USA) equipment using the commercial kit ECUGEN SARS-CoV-2 RT-qPCR (Starnewcorp-UDLA, Ecuador), a triplex assay based on the CDC protocol that includes the viral targets N1 and N2, and the RNase P gene as a control for RNA extraction quality. According to the manufacturer, its sensitivity is 97.7% and specificity 100% (Freire-Paspuel and Garcia-Bereguiain, 2021a; Freire-Paspuel et al., 2022). As we could not have access to a BSL3 lab for SARS-CoV-2 cultures, we could not address sample tittering with known concentration of SARS-CoV-2 viral particles. So, for viral loads calculation, the 2019-nCoV N positive control (IDT, USA) was used, provided at 200.000 genome equivalents/uL. This positive control is a plasmid including N1 and N2 viral gene targets sequences and it is a SARS-CoV-2 positive control recommended by CDC guidelines (Freire-Paspuel and Garcia-Bereguiain, 2021a; Freire-Paspuel and Garcia-Bereguiain, 2021b; Freire-Paspuel et al., 2022). Serial dilutions of the positive control were included in each set of samples RT-qPCR running, so an internal calibration curve with triplicates of known concentrations of genomic SARS-CoV-2 material was always available. A regression analysis was made for each of those calibration curves taking RT-qPCR Ct values for N1 and N2 targets and viral genomic material concentrations as variables; the equation obtained was used for viral load calculations for each set of clinical, finally expressed of an average of the values for N1 and N2 targets. Regression coefficients over 0.99 were obtained for the viral load calibration curves. The viral loads are expressed in copies/uL of RNA extraction, and the conversion factor for copies/mL of sample media is 200 in our experimental conditions.

The sensitivity, specificity, positive predictive value, and negative predictive value were calculated with confidence level of 95%. The t-Student was used to compare the Ct values obtained by each of those values for each protocol. Kappa-Cohen coefficient was calculated to measure the degree of agreement between two methods. All statistical analysis was carried out using SPSS Statistics 23 software.

A total of 388 nasopharyngeal samples were included in the study, of which 222 were positive and 166 were negative for SARS-CoV-2 using the column extraction method followed by RT-qPCR. 197 out of 222 SARS-CoV-2 positive samples tested also positive with the alternative homogenizer method. Moreover, 161 SARS-CoV-2 negative samples were also negative for the mechanical lysis method without RNA extraction (See Supplementary Table 1 ). So, the overall sensitivity was 88.74% (95% IC; 83.83 – 92.58) and specificity was 97% (95% IC; 93.11 – 99.01). Also, Positive Likelihood Ratio was 29.46 (95% CI 12.41 - 69.94); and Negative Likelihood Ratio was 0.12 (95% CI 0.08 - 0.17). Considering the SARS-CoV-2 infection rate obtained in the sample set included in this study, a positive predictive value of 97.52% (IC 95%; 93.3 – 99.6) and a negative predictive value of 86.56% (IC 95%; 81.6 – 91.5) was obtained. The Kappa Cohen value for the comparison between the RNA extraction method and the homogenization method was 0.844, which means a “very good” agreement between both protocols.

In Table 1 , the values of the clinical performance parameters for different Ct values thresholds are detailed. For Ct < 25, the sensitivity was 100%. For Ct < 30, the sensitivity was 97.92% (IC 95%; 96.01 – 99.83).

The analytical sensitivity for the homogenizer-based RNA extraction free method was addressed by comparing the Ct values obtained for the viral N1 and N2 targets ( Figure 2 ). The mean Ct value for N1 was 21.42 ± 7.10, and for N2 was 19.93 ± 5.8, with the column based RNA extraction protocol. The mean Ct value for N1 was 25.27 ± 4.81, and for N2 was 24.54 ± 5.25, with the homogenizer-based RNA extraction free method. These differences between both methods for the mean N1 and N2 Ct values were statistically significant (p<0.001). In Figure 2B , the linear regression analysis for the N1 and N2 Ct values for both protocols used is displayed, showing a significant linear adjustment with R² values of 0.63 and 0.46.

In Supplementary Table 2 , the Ct values and viral loads (RNA copies/uL) for the 25 false negative samples with the RNA extraction free protocol by homogenization are detailed. Only 4 of those false negative samples have N1 Ct values below 30, meaning viral loads over 10³ copies/uL. The sensitivity for the homogenizer-based RNA extraction free method for SARS-CoV-2 detection by RT-qPCR was 97.92% (189/193) for a Limit of Detection (LoD) of 10³ copies/uL.