Demographic and epidemiological characteristics of participants
Participants were aged 35–90 years, with a median age of 65 years of which 60% were men and 40% women. The median (interquartile range: IQR) age was 76 (63–86) years, being higher in women than in men, and 72.77% of patients were 65 years or older (Table 1).
Infection with SARS-CoV-2 was determined by RT-qPCR analysis of nasopharyngeal samples from patients of the study. From the same patients, we determined serum levels of IgG antibodies against the receptor binding domain (RBD) of the SARS-CoV-2 S1 spike protein, which are highly specific target of antibodies in SARS-CoV-2 patients (14, 15). As described, neutralizing anti-SARS-CoV-2 IgG antibodies are usually observed by day 9 after the onset of symptoms (16).
Once RT-qPCR and anti-Spike S1 IgG antibodies were determined, patients were divided into three groups: no infection (PCR negative, IgG negative) (n = 24), early infection (PCR positive, IgG negative) (n = 25) and acute/active infection (PCR positive, IgG positive) (n = 35) groups (Table 1).
Cytokine levels in participants
We then evaluated the serum concentration for the cytokines IL-6, TNF-α and IL-28B. Levels of IL-6 and TNF-α increased significantly in the acute phase of SARS-CoV-2 infection (Fig. 1), in good agreement with previous observations showing elevated IL-6 in the setting of severe Covid-19 disease (17, 18). In line with this, a positive association between those cytokines and the severity of the viral infection and mortality rate has been described (18). By contrast, we found that IL-28B notably decreased in most patients with COVID-19 (PCR positive) compared to controls (Fig. 1).
Cardiovascular biomarkers in participants
We then investigated the levels of the classical cardiac damage biomarkers cTnT and NT-proBNP in patient’s sera. Only 5 patients with acute infection had NT-proBNP levels higher than the cut-off value to predict the adverse outcome of severe COVID-19 (Supplemental Fig. 1), which was previously determined to be 88.64 pg/mL (19). No baseline cTnT elevations were detected in acute infection patients and only one patient in the early infection phase showed significant high levels of cTnT (Supplemental Fig. 1). In consequence, neither cTNT or NT-proBNP cardiac biomarkers significantly vary between control and PCR positive samples, which indicates that they were not useful to detect the early stages of infection.
We next sought to determine the levels of the novel biomarkers for cardiovascular events IMA and PAPP-A. IMA is a highly sensitive marker of hypoxia and detectable in the reversible early phase of myocardial ischemia, so we wonder whether it would be increased in SARS-CoV-2 infection, as ischemic stroke has been reported in patients with COVID-19 (20). Measurement of IMA levels in patient’s sera revealed a notably increase in early infected patients (PCR + IgG-) (Mean 94.92 ± 4.80 mg/ml) compared to controls (PCR - IgG -) (Mean 44.03 ± 4.35 mg/ml) (Fig. 2A). In the acute infection (PCR + IgG +), IMA levels (Mean 77.73 ± 3.25 mg/ml) were higher than controls, although not as high as in the early infection (Fig. 2A).
Receiver operator characteristic (ROC) curve analysis showed that the area under the curve (AUC) was 0.867 (95% CI: 0.775–0.959) for IMA determination in the total of patients, and AUC = 0.940 (95% CI: 0.881–0.999) for IMA determination in the early phase of SARS-CoV-2 (Fig. 2B), which indicates an excellent discrimination. The optimum diagnostic cut off point which maximized the sensitivity, and the specificity was determined to be 59.26 mg/mL, with a sensitivity of 90 % and a specificity of 75%. These results indicate that IMA determination may facilitate diagnosis of COVID-19 with relatively high sensitivity and specificity.
We further investigated the levels of the Pregnancy-associated Plasma Protein, PAPP-A, in the sera of COVID-19 patients. As shown in Fig. 3A, PAPP-A concentration of early infection group (Mean 5.618 ± 2.281 ng/ml) was above the levels of the control (Mean 0.258 ± 0.043 ng/ml) and the acute infection (Mean 0.231 ± 0.034 ng/ml) groups. It is worthy to note that 35% of COVID-19 patients in the early phase, had remarkably high levels (an average of 65-fold increase over controls) of PAPP-A (Fig. 3A). This result suggested that PPAP-A could be used as a biomarker for the early phase of COVID-19. This was confirmed by plotting the ROC curve and calculating the AUC for PAPP-A values of early infected patients. As shown in Fig. 3B, AUC = 0.801 (95% CI: 0.673–0.929) for early infected patients, indicating a very good discrimination. The best cutoff value for PAPP-A was 0.40 ng/mL, with a sensitivity and specificity of 61% and 86%, respectively. However, levels of PAPP-A of patients in the acute infection phase, were similar as controls and not useful for diagnosis as AUC of ROC curve was 0.526 (Fig. 3B).
The sensitivity, specificity, positive predictive value and negative predictive value of IMA, PAPP-A and combined IMA and PAPP-A for the different patient groups were calculated (Table 2).
The combined use of IMA and PAPP-A significantly improved the sensitivity, specificity, positive predictive value and the negative predictive value of Total COVID-19 patients to 93%, 75%, 39% and 98% respectively. AUC for the combination of IMA and PAPP-A for total COVID-19 patients was 0.90 (Fig. 4), which is higher than PAPP-A value and similar to IMA.
These results indicate that serum levels of IMA and PAPP-A could be used as new biomarkers for COVID-19. The notably increase of IMA and PAPP-A observed in the first stages of SARS-COV-2 infection should be considered when determining these biomarkers to diagnose other conditions.