SARS-CoV-2

Introduction 

SARS-CoV-2 is a novel beta coronavirus of likely zoonotic origin that emerged at the end of 2019 in Wuhan, China¹. SARS-CoV-2 differed from previous viral threats in the following manner:

a)     Marked transmissibility during the asymptomatic/very early symptomatic stage², unlike SARS

b)     Person-to-person transmission using airborne routes and also via fomites³, unlike Ebola,

c)     No previous immunity in the world population, unlike seasonal influenza,

d)     No known effective antiviral treatment (with regards to the first wave), unlike HIV, and

e)     No prior known effective vaccine  (with regards to the first wave). 

Due to the ‘perfect storm’ of all the above, SARS-CoV-2 rapidly spread and precipitated an international health crisis that has affected all countries. The associated mortality and morbidity has been a heavy burden. The mortality rate in hospitalized SARS-CoV-2 patients (excluding critical care) is 11.5% ( 95% CI 7.7 – 16.9) and 40.5% (95% CI 31.2 – 50.6) for critical illness⁴. Although frequently compared to influenza, there is no doubt that SARS-CoV-2 has greater risk of mortality (16.9% for hospitalized SARS-CoV-2 patients versus  5.8% in hospitalized influenza patients, with a relative risk of death of 2·9 (95% CI 2·8-3·0) and an age-standardised mortality ratio of 2·82)⁵.

Prognostic scores are important to maximize survivorship and minimize morbidity. Sophisticated scoring systems have been proposed but have not performed consistently ^6-9. We sought an easily measurable, dependable single-parameter biomarker to predict mortality in swab-positive SARS-CoV-2 patients. Interestingly, our meta-analyses have revealed distinct differences in the effectiveness of different biomarkers in different regions of the world. For example, admission CRP levels are a good prognostic marker for mortality in Asian countries, with a pooled AUC (area under curve) of 0.82 (95% CI [0.79, 0.84]) from 19 studies, but only an average predictor of mortality in Europe and North America, with a pooled AUC of 0.66 (95% CI [0.62, 0.71]) from 18 studies (P<0.0001, Fig. 1).

We have mapped the root studies on this website so that all may see which biomarkers perform well in their locale.

We see the same pattern for D-dimer and IL-6 – they are good predictors of mortality in Asian countries (pooled AUCs of 0.78 [0.74, 0.81] and 0.86 [0.78, 0.91] respectively) but not in Europe and North America (pooled AUCs of 0.69 [0.66, 0.72] and 0.71 [0.64, 0.77] respectively; P<0.001 for both compared to Asian counterparts). This has significant implications and explains why the prognostic scores that are being proposed for SARS-CoV-2 do not perform evenly in different countries. 

There are multiple reasons why this might be true. This might be due to genetic differences and the fact that the Asian cohorts are younger than the European/N. American cohorts in all five parameters we investigated (CRP, D-dimer, troponin, urea, and IL-6). It is possible that the principal mode of death is different in Asia, where younger patients are dying from cytokine storm, while in Europe older people are dying from multi-organ failure. It is also possible that there is a ‘training’ effect, with the West having had prior warning and the Asian experience to buffer the storm. 

We did find two biomarkers that performed well in all cohorts. Both urea and troponin had pooled AUCs in excess of 0.77 regardless of geographical distribution. This implies that end-organ damage at the time of presentation is a key prognostic indicator of severity. 

We have mapped the root studies on this website so that all may see which biomarkers perform well in their locale. We also propose a free-to-use program that the healthcare community can use to check whether their biomarker of choice is effective in their population. We would request that results from this be uploaded so we can periodically update the website.