Herein we report that plasma SARS-CoV2 viral loads, but not ETA viral loads, are associated with prolonged mechanical ventilation, reduced hospital-free days, and higher risk of VAP. Previous studies demonstrated an association between severity of illness in hospitalized patients and high SARS-CoV2 plasma RNA4,9 and nucleocapsid antigen levels10-12. Our study adds to the SARS-CoV2 literature by being the first to show that this relationship extends to ventilator complications as well. While a prior study found the SARS-CoV2 plasma viral loads were proportional to ETA viral loads 9, we found no such association.
The relationship between SARS-CoV-2 respiratory tract VL and disease severity has been inconsistent. The majority of the studies measured VL in the upper airway (via nasal or nasopharyngeal swabs) and reported significantly higher upper airway VL in patients with greater symptoms burden1 or death4,6. Other studies reported that a lower RT-PCR cycle threshold (Ct) (indicating higher viral load) predicted increased disease severity and death2,3,5. Conversely, while Knudtzen et al. found an association between low Ct and morbidity in hospitalized patients, Ct did not predict admission to the hospital2. Similarly, other studies found no relationship between Ct in outpatients and those admitted to the hospital or ICU13,14. One of the largest studies originating from Bahrain studied 1057 patients across multiple centers and found no association between Ct values and disease severity15. These conflicting results suggest that such studies need to be interpreted with caution. da Silva argues that Ct cannot be directly compared across studies given differences in RT-PCR assays, and exclusive use of Ct over VL may represent a statistical bias16. These data also suggest respiratory viral load sampling may not be consistently reliable to gauge viral burden. Our study is among the few to measure VL in the lower respiratory tract but also found no association between ETA VL and clinical or ventilator outcomes within 90 days of test positivity. In contrast to our study, Buetti et al. performed a multicenter study of 90 subjects with COVID-19 ARDS and found survivors displayed a significantly lower respiratory tract VL compared with non-survivors after 35 days 17. The immediate reason for the discrepant findings is not clear; it could relate to differences in patient population, ETA acquisition method or timing, PCR methodology, other unmeasured variables, or a type 1 error.
The inconsistent association between upper or lower respiratory tract VL and Covid-19 morbidity/mortality across studies suggests this sampling technique may be limited by variability. Plasma sampling, however, is generally consistent across studies and may be a more reliable manner to measure viral load. It is also possible that viral shedding in the respiratory tract does not equate to infectivity. As all of the studies use RT-PCR rather than viral culture, it is not known whether the virus detected by PCR represents viral particles or live, infectious pathogen. In critically ill patients, viral shedding in the lower respiratory tract is known to be more prolonged, as high as 50-60 days post-infection, and in higher quantities compared to that of the nasopharynx and throat 6,8. As such, it is possible that these assays are detecting noninfectious viral particles.
Our study has a number of limitations. One is the small cohort size and single-center design. Second, both plasma and ETA samples were obtained on the same study days, but were not necessarily the same ICU days for each subject, which confounds comparison between subjects. Third, we used ETA, which is highly variable among subjects and dependent on the amount of secretions available, whereas a bronchoalveolar lavage might be a more sensitive test for lower respiratory tract infection. Fourth, all of the samples for this study were collected prior to the Omicron wave, and may not be reflective of the current Covid-19 infections due to the new subvariants.
In conclusion, in critically ill patients admitted to the ICU with COVID-19 respiratory failure, elevated plasma viral load was associated with increased days on mechanical ventilation, incidence of ventilator-associated pneumonia, and increased hospital length of stay. ETA viral load was not associated with in-hospital complications or mortality.