Multiplex Assay Setup

On the basis of diagnostic testing during the SARS-CoV-2 pandemic (8; C. Manohar et al., unpub. data, https://doi.org/10.1101/2021.10.13.21264919), we chose a dual-target approach, in which 1 assay targets a conserved sequence of the Orthopoxvirus genus, not including variola major or minor viruses (9), and the other targets a MPXV-specific sequence (10) (Table 1). The cobas 5800, 6800, and 8800 systems (Roche Diagnostics, https://diagnostics.roche.com) use a spike-in RNA full process control that is added automatically during extraction. The corresponding internal control assay is preloaded in the open channel reagent for use with cobas omni Utility Channel (Roche Diagnostics) (Table 2). We modified and optimized all assays for use on cobas 5800, 6800, and 8800 systems, including 2′O-methyl-RNA-modified primers and internal quenchers for TaqMan probes, as previously described (11).

In Silico Evaluation

As part of a support request for Utility Channel applications, we submitted all sequences of the duplex assay to Roche Diagnostics for evaluation of inclusivity and potential primer-probe interactions. The submitted sequences were aligned to currently available MPXV and orthopoxvirus sequences available in public databases.

Analytical Performance Evaluation

We conducted technical performance evaluations for the assays according to new European Union regulations (Regulation 2017/746 EU IVDR, https://euivdr.com). For reference material, we used inactivated cell culture supernatant containing MPXV recovered from a clinical case in central Africa in 1987 (12) and inactivated modified vaccinia virus Ankara (MVA) SARS-CoV-2 vaccine. To obtain a quantitative MPXV standard, we purified nucleic acids by using a MagNA-pure96 extractor and Viral NA Small Volume Kit (Roche Diagnostics) and analyzed on a QIAcuity digital PCR instrument (QIAGEN, https://www.qiagen.com) in conjunction with 3 different previously described quantitative PCRs (qPCRs): 1 for nonvariola (NVAR) orthopoxviruses (9); 1 for MPXV (10); and the research-use only (RUO) LightMix Modular Orthopoxvirus Assay (TIB MOLBIOL, https://www.tib-molbiol.de).

We determined the lower limit of detection by serial 2-fold dilution of MPXV standard in universal transport medium from 100 to 0.78 copies/mL and 21 repeats per dilution step. Using MedCalc statistical software (https://www.medcalc.org), we determined the limit for 95% probability of detection. We assessed linearity by 10-fold serial dilution of MPXV standard (5 repeats per dilution step) at concentrations of ≈10¹–10⁷ copies/mL. We determined linearity and intra-assay variability by using Validation Manager software (Finbiosoft, https://finbiosoft.com). Concentrations represent copies per mL of specimen. For empirical inclusivity and exclusivity testing, we used the assay to test a set of 53 samples, including clinical samples, reference material, and external quality controls from a range of bloodborne and respiratory pathogens (Appendix Table 1). We used an experimental MVA vector-based SARS-CoV-2 vaccine as reference material for a non-MPXV orthopoxvirus.

Clinical Evaluation and Follow-Up Samples

For clinical validation, we used the RUO LightMix Modular Orthopoxvirus assay as reference test, which we performed according to manufacturer’s recommendation by using the MagNA-pure96 system with 200-µL extraction volume. In total, we tested 67 clinical samples consisting of respiratory, skin, and genital swab samples with both assays. Of those samples, 11 were positive for MPXV DNA, which we obtained from 2 confirmed clinical cases in Hamburg, Germany. We analyzed 33 consecutive clinical samples from the same 2 patients and 2 additional cases by using the duplex assay (Appendix Table 2). The clinic provided globalized patient characteristics.

In Silico Analysis

We did not detect any concerning oligo interactions (Appendix Figure 1). Target-1: NVAR was still a 100% match for all but 1 MPXV sequence, which had 1 low-risk mismatch. NVAR also had high sequence similarity with many other orthopoxviruses but might not be optimal for reliable detection of camelpox or cowpox (Appendix Figure 2). Target-2: MPOX is a perfect match for almost all Congo Basin strain MPXV sequences but has a known mismatch for West Africa strain sequences in the probe region. This mismatch is expected to slightly reduce relative fluorescence increase signals, as demonstrated in the clinical sample set. Other orthopoxviruses have extensive sequence mismatches with this assay and are not expected to produce detectable signals (Appendix Figure 3).

Analytical performance

Figure 1

Figure 1. Linearity data for the dual-target monkeypox virus assay rapidly adapted from established high-throughput molecular testing infrastructure. A) Nonvariola orthopoxvirus target; B) monkeypox virus target; C) absolute Ct for nonvariola orthopox...

We determined LoD was 9.697 (95% CI 7.424–15.327) copies/mL for the NVAR assay and 6.359 (95% CI 4.908–10.110) copies/mL for the MPOX assay by probability of detection analysis. Overall LoD for both targets combined was 4.795 (95% CI 3.598–8.633) copies/mL. We compiled hit rates (Table 3) and probability of detection plots (Appendix Figure 4) for the assay. The assay showed excellent linearity. Cycle threshold (Ct) values were 37–18, ≈10¹–10⁷ copies/mL, and pooled SD and 95% CI were within linear range: Ct 0.194, SD 0.0662% for NVAR; Ct 0.175, SD 0.618% for MPXV (Figure 1).

No false positives occurred within the inclusivity-exclusivity set. The MVA vector vaccine was correctly detected by the NVAR assay, and not by the MPXV assay (Appendix Table 1).

Clinical Evaluation

Figure 2

Figure 2. Amplification curves of clinical samples, including internal controls for dual-target monkeypox virus assay rapidly adapted from established high-throughput molecular testing infrastructure. A) Nonvariola orthopoxvirus; B) monkeypox virus; C) internal control....

In total, we tested 67 clinical samples, consisting of respiratory, skin, and genital swab samples, with both assays. Of those, 11 samples obtained from 2 confirmed clinical monkeypox case-patients in Hamburg, Germany, were positive for MPXV DNA. We noted 100% positive (11/11) and 100% negative agreement (56/56) for the 2 assays (Figure 2).

Results from Different Sample Types and Timepoints

Another 33 clinical samples were longitudinally collected from 4 patients, all of whom were male, 20–40 years of age, and had 6–50 skin lesions; 1 patient had known HIV infection under treatment. Lesion swabs generated Ct values of 13.3–16.1, oropharyngeal swabs Ct values of 13.1–33.3, and blood samples Ct values of 30.3–38.4. A small sample set of urine had only low concentrations of viral DNA, Ct 31.1–37.8. A single patient provided seminal fluid, which had Ct values of 32.9 for the NVAR assay and 33.9 for the MPXV assay when diluted in guanidine hydrochloride solution (Appendix Table 2).