Sensitivity of three commercial tests for SARS-CoV-2 serology in children: an Italian multicentre prospective study

The Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) was declared a global pandemic in March 2020 by the World Health Organization (WHO) [1, 2].

At present, the standard diagnostic confirmatory test for Coronavirus Disease (COVID)-19 is based on the detection of nucleic acids of SARS-CoV-2 by nucleic acid amplification tests on respiratory samples [1]. Antigen tests and rapid molecular-based tests are an alternative, being suitable for use as point of care [3].

Despite the limited role in the diagnosis of acute infection, serological tests are important for surveillance purposes and epidemiological assessment of the immunization status of the population [4]. Moreover, serology is a cornerstone in the definition of cases of Multisystem Inflammatory Syndrome in Children (MIS-C), when the clinical presentation is suspected. Another possible use is the assessment of COVID-19 vaccine immune responses and durability [4, 5].

Although the humoral response to SARS-CoV-2 currently is incompletely defined, it appears that approximately 60% of infected individuals produce IgM antibodies about 4 days post-symptoms onset, with a peak between 14 and 21 days and then decline. IgG levels begin to rise at about 7–14 days, peaking at around day 25 [6]. It is unclear how long IgG levels are sustained although some individuals has detectable IgG antibodies at least 6–7 months after onset [7, 8]. Moreover, the evaluation of IgA levels in a larger number of COVID-19 patients is still lacking [9, 10].

The US Food and Drug Administration (FDA) has issued Emergency Use Authorizations for hundreds of serological assays to support COVID-19 diagnosis, and at present more than 1 thousand immunoassays are either commercially available or in development [11, 12]. According to the WHO indication, ≥ 95% and ≥ 97% as acceptable criteria for sensitivity and specificity, respectively [13]. Test types include formal laboratory-based assays such as enzyme-linked immunosorbent assays (ELISA), chemiluminescent immunoassays (CLIA) and point-of-care rapid lateral flow immunoassays (LFIA) [1, 4]. Despite the large number of available serology kits, studies on serological tests in children with SARS-CoV-2 infection are limited.

The main objective of this study was to evaluate the sensitivity of three widely utilized SARS CoV-2 serology commercial assays, Diesse Diagnostics (IgG, IgA, IgM) and Roche Diagnostics, both Roche N (IgG) and Roche S (IgG) assays. The secondary aim was to perform a head-to-head comparison of the diagnostic sensitivity of those three assays.

This multicentre prospective study describes the test performance of three anti-SARS-CoV-2 assays on 149 children with rRT-PCR confirmed SARS-CoV2 infection. To the best of our knowledge, this is the first study comparing the performance of three widely utilized SARS-CoV-2 serology commercial assays in children. The main finding of our study is the high sensitivity of CLIA assays (Roche S and Roche N) after 14 days since symptoms onset. Moreover, CLIA tests (Roche N and Roche S) were found to be significantly more sensitive than ELISA test (Diesse). In fact, a low sensitivity was found for Diesse IgA, IgM and IgG in our study, and the lowest sensitivity was found for Diesse IgG in all the time frames.

Few studies are available on SARS-CoV-2-specific IgA [9, 10]. The addition of IgA to serological tests analyzing IgM and IgG could increase the sensitivity of SARS-CoV-2 diagnosis in the early stage of infection [15]. In fact, the involvement of the upper airways, highly containing mucosal immune cells, could explain IgA production. In a study by Chiereghin et al. the sensitivity of ELISA IgA was 0.84 and significantly improved overtime in symptomatic patients, resulting in an overall sensitivity of 0.94 [16]. On the contrary, according to our results, Coste et al. demonstrated an insufficient performances of IgA and IgM for the diagnosis of COVID-19 [17]. Furthermore, some studies reported a cross-reactivity of ELISA IgA with other respiratory viruses such as influenza A and B and with other human coronaviruses [18‐20].

In our study, IgM and IgA detected by ELISA performed better in the time interval 15–28 days from symptom onset while, Diesse IgG had a low sensitivity in all the time intervals. In general, the sensitivity of IgG and IgM based tests was low in the first week (1–7 days) of symptom onset and high in the third week or later (> 14 days) [1]. A metanalysis by Vengesai et al. showed higher pooled sensitivity with IgG and IgM based ELISA tests of 0.83 and 0.84, respectively [1]. Similarly, other three meta-analysis observed similar pooled sensitivity ranging from 0.70 (95% CI 0.55–0.84) to 0.84 (95% CI 0.70–0.92) [21‐23]. The lower sensitivity of ELISA IgG found in this study compared to literature could be related to the small number of patients tested after 14 days since symptoms onset [1, 24, 25]. Moreover, the difference between literature data and our results on ELISA sensitivity could have few other explanations. In fact, it could be also be attributed to the targeted SARS-CoV-2 antigens contained in each kit and to the ELISA commercial kit used, considering also that the cut-off might also play a role [19]. Another important difference is that the available literature is limited to the adult population. The dynamics of the antibody response has been well described in adults [26], while there are few data in the pediatric population. A difference in the distribution, maturation and functioning of viral receptors has been mentioned as a possible reason for the age-related peculiarities [27]. In a in a Spanish multicenter study on 324 SARS-CoV-2 rRT-PCR positive children, 24% of them failed to seroconvert after the infection and patients with mild disease and shorter time to rRT-PCR negativity seroconverted less often than patients with more severe disease and who had more prolonged rRT-PCR positivity [26]. Moreover, studies of MIS-C suggested that not all patients developed detectable antibodies despite a strong dysregulated immune response [28, 29]. The evaluation of tests sensibility concerning the disease severity and special conditions such as MIS-C was not performed in our study, due to the limited number of cases.

Overall, CLIA tests exhibited a better sensitivity in our study population, with a better performance of Roche S in the interval 15–28 days. The sensitivity of Roche S and Roche N has been also evaluated in another study on adults, showing a sensitivity of 0.96 (95% CI 0.92–0.98) and 0.92% (95% CI 0.87–0.96), respectively [30].

We found that Roche S had a higher sensitivity at 15–28 days and Roche N at 28–84 days. This trend toward higher sensitivity over time is in accordance with previous studies [31].

The results in the literature are controversial. In a French study on 68 patients between 7 and 81 years of age eight commercial assays based on CLIA, ELISA and enzyme-linked fluorescent assay (ELFA) technologies were compared [32]. In this study, Wantai ELISA showed the best sensitivity, whereas Liaison CLIA the worst one [32]. In another study by Wolff et al., the highest overall sensitivity among the examined methods (Roche N CLIA, Liaison CLIA and Euroimmun ELISA) was achieved by using Euroimmun ELISA with a combined detection of IgG/IgA (0.86, 95%CI 0.78–0.92) [33].

In contrast, some studies demonstrated the superior performance of CLIA-based technique over ELISA-based ones, similarly to our study [34, 35]. In a study by Schnurra and colleagues comparing seven commercial antibody tests including Euroimmun ELISA, Siemens CLIA and Roche CLIA, the highest sensitivity was obtained by Siemens antibody testing followed by Roche and Euroimmun [34]. Other studies revealed that the diagnostic performance of CLIA tests is comparable to ELISA [10, 36‐40]. In an Italian study on 184 serum samples from 130 COVID-19 patients and 54 SARS-CoV-2 negative subjects, four CLIA assays (Abbott SARS-Cov2 IgG, Roche N, Ortho SARS-CoV-2 total and IgG) and one ELISA (Diesse ENZY-WELL SARS-CoV-2 IgG) assay were compared [40]. The overall sensitivity of Roche N CLIA and Diesse ELISA was 0.78 (95% CI 0.70–0.85) and 0.83 (95% CI 0.75–0.98), respectively. A higher sensitivity was reached after 12 days since symptom onset. Another study by the same authors suggested that IgG levels measured by Maglumi CLIA and Euroimmun ELISA assays were comparable and the clinical agreement between these methods was 0.90 [10]. In another study by Egger et al., SARS-CoV-2 antibodies were measured with the Elecsys assay (Roche N CLIA) and the Edi ELISA in 64 patients, showing a sensitivity of 1.00 for Roche and of 0.94 for Edi ELISA 15–22 days after symptom onset [38].

Some studies suggested that combining N- and S-based tests may enhance true positivity and can be beneficial when extremely sensitive antibody tests are not accessible [30, 34]. However, in our study, the combination of tests did not significantly increase the sensitivity compared to the use of a single test (except for Diesse), nor did a combination performed better than the others. Similarly, in the study by Andrey et al., either Roche S alone or Roche S and Roche N parallel testing (either one positive leading to a positive result) displayed a sensitivity of 100% [30].

Moreover, a secure benefit of combining serology using both anti-N and anti-S antibody detection would be the capability of differentiating antibodies induced by SARS-CoV-2 infection (with both anti-S and anti-N antibodies) versus vaccine-induced antibodies (only anti-S antibodies) [30]. This application would be more useful in view of the recent authorization of anti-SARS-CoV-2 vaccines also for the pediatric age [41, 42].

The main limitation of our study is the small number of children enrolled. Moreover, the three assays were not available for all the enrolled patients due to the insufficient blood samples obtained in some cases, especially in young children. Another limitation is that the serology was not done at standardized timing but on the occasion of blood tests performed during hospitalization. This point led to the fact that in our study the median timing of serology was of 7 days, whereas the best sensitivity was reached after 14 days since symptoms onset. Therefore, data regarding serology after 14 days since symptoms onset were limited to a few patients. Finally, data on serology at follow-up were not available in this study.

Serological tests could play a role in the diagnosis and follow-up of children with SARS-CoV-2 infection, and CLIA tests seem to better perform in this population. Large prospective studies are needed to better define the characteristics of those tests in the pediatric population.