Diagnostic Performance of Ag-RDTs and NAAT for SARS-CoV2 Identification in Symptomatic Patients in Catalonia

The use of rapid antigenic tests (Ag-RDTs) to diagnose a SARS-CoV-2 infection has become a common practice recently. This study aimed to evaluate performance of Abbott PanbioTM Ag-RDTs with regard to nucleic acid amplification testing (NAAT) in the early stages of the disease. A cohort of 149,026 infected symptomatic patients, reported in Catalonia from November 2020 to January 2021, was selected. The positivity rates of the two tests were compared with respect to the dates of symptom onset. Ag-RDTs presented positivity rates of 84% in the transmission phases of the disease and 31% in the pre-symptomatic period, compared to 93% and 91%, respectively, for NAAT. The detection of many false negatives with Ag-RDTs during the pre-symptomatic period demonstrates the risk of virus dissemination with this diagnostic technique if used outside the symptomatic period.


Introduction
Currently, due to its high sensitivity and specificity, nucleic acid amplification testing (NAAT) has become the gold standard for the detection of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [1]. Ag-RDTs enable sensitive detection of high viral loads (Ct values ≤25 or >10 6 genomic virus copies/mL) which usually arise in the presymptomatic (1-3 days before symptom onset) and early symptomatic phases of infection (within the first 5-7 days of infection) [2,3].
Although sensitivity is lower, especially in asymptomatic patients, Ag-RDTs offer advantages in terms of speed and low-cost strategies compared to NAAT [4,5].
In December 2020, the European Union adopted a common response for the use, validation and recognition of Ag-RDTs as a diagnostic method to detect SARS-CoV-2 infection [6]. The European Centre for Disease and Control (ECDC) agreed on a selected list of validated Ag-RDTs along with recommendations on their use for the diagnosis of COVID-19, based on the best available evidence to date, from systematic research of clinical trials [4]. Panbio TM COVID-19 Ag Rapid Test Device (Abbott Diagnostics, Jena, Germany) started to be used massively in Catalonian primary care centers and hospitals at the end of October.
Although indications for the use of Ag-RDTs were limited, a mitigation strategy was implemented in order to relieve pressure on overburdened laboratories [7]. To overcome the bottleneck of NAAT testing, the use of Abbott Panbio TM Ag-RDT was widespread

Results
Between 1 November 2020 and 31 January 2021, 219,138 patients received a confirmed diagnosis of COVID-19. Among them, 149,026 were symptomatic (68%). Demographic and epidemiological characteristics are shown in Table 1.
During the studied period, 7-day cumulative incidence rates (cases per 100,000 inhabitants) varied between 101 and 350, being above 150 for 9 weeks (from 1 November to 15 November and 14 December to 31 January) and below 150 for 4 weeks (from 16 November to 13 December).
Positivity rates were 80% and over between +1 and +7 days from the symptom onset for Ag-RDTs and between −2 and +10 days for NAAT. The highest probability of detecting a positive case was on the third day after symptom onset (93.0% of positive tests) for Ag-RDTs and on the first day after symptom onset for NAAT (95.5%) ( Table 2). During the transmission period, from −2 to + 7 days from symptom onset [10], the overall positivity rate was 84.2% for Ag-RDTs and 93.1% for NAAT (p-value < 0.001) ( Table 3).
The highest difference in positivity rate between the two tests was found in the presymptomatic period, the second day before symptom onset being the day with the highest difference ( Figure 1; Tables 2 and 3). During the two days before symptom onset, the positivity rate was 30.8% for Ag-RDTs and 90.9% for NAAT (p-value < 0.001). Differences between the two tests were higher in the 15-44 age group (28% for Ag-RDTs vs. 91% for NAAT), and lower in the 0-14 age group (40% for Ag-RDTs vs. 95% for NAAT) and in patients with severe outcomes (36% vs. 83% in hospitalized patients and 39% vs. 82% in exitus patients) than in patients with milder symptoms (Table 3, Figure S1, Table S1).
In health workers and residents of nursing homes the use of Ag-RDTs was limited with respect to NAAT (Table S1), and positivity rates for Ag-RDTs were lower than in the rest of patients during the transmission period (79.9% in health care workers and 74.9% in residents of nursing home) and especially during the pre-symptomatic period (22.3% in health care workers and 24.4% in nursing home residents) ( Table 3).    Pre-symptomatic and symptomatic In higher incidence conditions, greater performances by both Ag-RDTs and NAAT were observed, especially in Ag-RDTs during pre-symptomatic period and post-transmission period (Table 3).
From the eighth day after symptom onset, the positivity rate of Ag-RDTs began to drop below 80% and from the 12th day to drop below 30%, whereas in NAAT it remained higher than 70% until the 14th day (Figure 1). Differences in drop rate were observed after the 12th day in all categories for those under 65 years old and in health workers; for those over 65 who were hospitalized or died, the Ag-RDT rate remained above 50% until the end of the observed period (Table 2, Figure S1).

Discussion
According to current guidelines on the use of Ag-RDTs [11], authorized Ag-RDTs, and especially Abbott Panbio TM , have reported consistently high levels of specificity, but they do not achieve equal levels of sensitivity compared to NAAT [12,13].
Due to the evidence of SARS-CoV-2 transmission in pre-symptomatic and asymptomatic patients [14,15], results presented in this study support the statement regarding a notably higher reliability of NAAT compared to Abbott Panbio TM Ag-RDTs during presymptomatic phases. Moreover, the high number of false negative cases detected in our study with Ag-RDTs during the pre-symptomatic period indicates that in suspected patients with no symptoms the negative antigen result should be confirmed with a NAAT [16]. Additionally, results from molecular assays during post-symptomatic period should take into account the possible presence of a residual SARS-CoV-2 RNA load (quantification cycle value greater than 30) with a subsequent negative Ag-RDT result, which could be pointed as a confounder of "real positive" cases with ongoing virus replication [17,18].
Our results showed an accurate performance of Ag-RDTs in the age-group 5-14, unlike other studies [19,20]. The differences between age-groups could be related to differences in SARS-CoV-2 viral load [20,21]. However, age-related differences in the performance of Ag-RDT remain unknown. These results highlight the need to assess whether the widespread use of Ag-RDTs in pediatric patients can help to prevent and control the covid-19 pandemic. For this reason, more research in the clinical assessment of Ag-RDTs being used on patients of extreme age is needed.
Furthermore, we found greater performance of Ag-RDTs in a higher prevalence setting, especially from 8th day after symptom onset, probably due to higher viral load associated with higher virus dissemination [22].
In this sense, it should also be noted that the higher Ag-RDTs performance from the 7th day after symptom onset in adults aged 65 and over and especially in hospitalized patients and the deceased could be explained by the higher duration of infectious viral shedding in those groups. Previously published studies indicated that the presence of a high viral load of SARS-CoV-2 on admission was associated with a higher risk of respiratory failure and mortality [23,24]. It is, therefore, necessary to take into account that the duration of infectious viral shedding currently described is longer for severely-ill patients, which explains the higher positivity rates of NAATs and RDTs over time in this group [25].
Nevertheless, this study presents limitations: Assumptions about date of symptom onset may be affected by recall bias and uncertainty due to inaccurate recall of the symptoms onset by the study subjects [26]. Comparison of the cycle threshold (Ct) values of RT-PCR was not possible due to the lack of data availability in the existing databases. Furthermore in this respect, the high variability between analytical interpretive methods and laboratory factors must be taken into account [27]. Moreover, results should be taken with caution as the present research does not apply a single-subject design and the two techniques were not replicated in parallel for all patients. For this reason, this study was not intended to provide estimates of sensitivity and specificity, but to evaluate the appropriateness of the use of both testing procedures in a pandemic containment strategy.
The use of Ag-RDTs in pre-symptomatic close contacts of positive SARS-CoV-2 cases is a common practice in Catalonia. In the last few months, XVEC has observed some relaxation in the compliance of quarantine measures, usually by individuals who obtained negative Ag-RDT during the asymptomatic phase (and who are likely to end up with a positive NAAT). This misbehavior, in addition to the possible high number of Ag-RDT false negatives, leads us to think that the use of this type of test could even be counterproductive in certain scenarios, leading to a possible increase in the spread of the virus in the community.
Appropriate use of Ag-RDTs included scenarios with a high expected prevalence of disease, when NAAT capacities were not available and receiving timely results is criticalfor example, for contact tracing purposes [4]. In this sense, the studied period presented a 7-day incidence rate of over 100 cases per 100,000 inhabitants and a NAAT positivity rate of over 5%. The difficult situation in Catalonia in that period and the saturation suffered by the laboratories justifies the choice of including Ag-RDTs in the containment strategy by public health authorities.
Findings from this paper do not aim to support nor to retract Ag-RDTs suitability, but to limit its use to the largest possible extent, by only and exclusively setting its performance as a response to scenarios that are clearly recommended by the scientific community and Public Health Agencies.