We collected 21 pairs of EDTA whole blood and serum specimens from 21 patients with MERS-CoV after admission to the National Medical Center in Seoul, South Korea. MERS-CoV infection initially was diagnosed by the Korea Centers for Disease Control and Prevention using respiratory specimens (11). After admission, each patient was reassessed for epidemiologic information and clinically managed with monitoring.

Specimens were stored at −80°C before analyses. Viral RNA was extracted and eluted with a MagNA Pure LC 2.0 automated nucleic acid extractor and MagNA Pure LC total nucleic acid isolation kit (both from Roche Diagnostics, Mannheim, Germany) according to the manufacturer’s instructions. Specimen volumes were 100 μL EDTA whole blood and 200 μL serum, and elution volumes were 100 μL for EDTA whole blood and 50 μL for serum. One-step, real-time reverse transcription PCR (rRT-PCR) was performed for the 3 MERS-CoV gene regions (upstream envelope [upE], open reading frame [ORF] 1a, and nucleocapsid) with an AgPath-ID One-step RT-PCR kit (Applied Biosystems, Foster City, CA, USA) and an ABI7500 real-time PCR system (Applied Biosystems). Human ribonuclease (RNase) P was amplified in parallel for sample quality control (5,12,13). In each test, the viral RNA was considered detected when amplification before cutoff was observed from at least 2 different targets in MERS-CoV with pass of sample quality control. The robustness of rRT-PCR was demonstrated for qualitative concordance of positivity or negativity by using different types of specimens (Technical Appendix Table 1). We calculated the viral copy concentration in the blood using standard curves constructed from the cycle threshold (C_t) of serially diluted 10⁵ copies/μL of upE RNA (provided by the University of Bonn Medical Center, Bonn, Germany).

The results of the blood viral RNA analyses did not affect clinical management. We assessed the relationships among clinical and molecular diagnostic factors with the IBM SPSS Statistics program 22.0 (SPSS Inc., Chicago, IL, USA). In each test, p<0.05 was considered statistically significant. The institutional review board of the National Medical Center approved this study (H-1508-057-002).

We assessed patient demographics, their clinical features, and disease outcomes (Tables 1, 2; Technical Appendix Table 2). The time difference was an average of 1.5 days (median 1, range 0–5 days) between when the initial diagnostic respiratory specimens and blood specimens were obtained. At admission, viral RNA was detected in 6 (29%) of 21 EDTA whole blood and 6 (29%) of 21 serum samples from infected patients. Two patients showed viral positivity in either specimen subtype of EDTA whole blood or serum; therefore, the overall detection rate for MERS-CoV was 33% (7/21) in blood. The concordance rate of viral assay was 90% (19/21) between EDTA whole blood and serum specimens. Blood virus concentration was 3,130 copies/mL EDTA whole blood (range 2,080–17,400 copies/mL), equivalent to a median upE C_t of 37.55, and 1,300 copies/mL serum (range 490–10,200 copies/mL, median upE C_t 36.88).

Blood viral RNA positivity at admission was associated with fever >37.5°C on the sampling date (p = 0.007), requirement for mechanical ventilation during the following clinical course (p = 0.003) and extracorporeal membrane oxygenation (p = 0.025), and patient death (p = 0.025, all by 2-tailed Fisher exact test; Figure). Blood viral RNA positivity was not associated with viral C_t in the initial diagnostic lower respiratory specimens, or requirement of oxygen supplementation during the following clinical course. Between the blood viral RNA-positive and -negative groups, we found no differences in age, duration from symptom onset to diagnosis of MERS-CoV infection, or an invasive procedure before the specimens were obtained (Technical Appendix Table 3).

Figure

Figure. Differences in survival among Middle East respiratory syndrome coronavirus–infected patients, South Korea, 2015. A, B) Survival difference between the blood viral RNA-positive (solid line) and -negative (broken line) groups. Survival was...

Viral loads in the lower respiratory specimens at the initial confirmatory diagnosis showed no effect on patient survival (Figure). Patient death was not associated with length of time from symptom onset to diagnosis of MERS-CoV infection (Technical Appendix Table 3).

Our results showed that the detection rate of blood viral RNA was low in the early phase of infection in patients with a confirmed diagnosis, similar to results from a previous study (14). These findings contrasted with those of severe acute respiratory syndrome coronavirus infection (15). Therefore, in the case of MERS-CoV infection, blood does not have the highest diagnostic yield for the initial confirmatory diagnosis. The viral load in blood was low, even in detected cases. A proportion of MERS-CoV isolates in the 2015 Korea outbreak harbored a C→T substitution in the third nucleotide of the ORF1a primer binding site (GenBank accession nos. KT374052–374055). This mismatch may partially contribute to the insensitivity of ORF1a assay observed in this study. An alternative sensitive target replacing ORF1a might be useful in studies using blood specimens.

Blood viral RNA has been detected in a few case reports of MERS-CoV fatalities (8–10). Our data of 42 specimens from cross-sectional time points focusing on early viremia showed that blood viral RNA was present in a subpopulation of patients and that these patients had significantly poorer prognoses, as demonstrated by the need for more frequent mechanical ventilation and the increased risk for death. Further large studies using serial daily specimens that are collected throughout the admission period–both upper and lower respiratory specimens and paired measurement of viral RNA and antibody in blood–might help overcome the limitation of the current study, which included relatively small numbers of deceased patients (24% [5/21]).

Our data showed a detection rate of 33% for viral RNA in blood at initial diagnosis, which was insufficient for initial confirmatory diagnosis. Blood viral RNA at the early phase was related to a worse clinical course in infected patients and might be a good prognostic indicator of severe outcome. Measuring blood viral RNA at hospital admission might be useful.