Impaired performance of SARS-CoV-2 antigen-detecting rapid diagnostic tests at elevated and low temperatures

Highlights • Analytical sensitivity of SARS-CoV-2 Ag-RDTs ranges from 1.0 × 106-5.5 × 107 copies/mL.• Even short-term exposure to 37 °C reduces sensitivity of SARS-CoV-2 Ag-RDTs.• Elevated temperatures impair sensitivity at clinically relevant virus concentrations.• Low temperatures limit SARS-CoV-2 Ag-RDT specificity.• Storage and operation of SARS-CoV-2 Ag-RDTs at recommended conditions is essential.


Introduction
Advantages of SARS-CoV-2 antigen-detecting rapid diagnostic tests (Ag-RDTs) include fast results and their applicability on site without dependence on laboratory settings. With a constantly growing number of commercially available Ag-RDTs on the global market, the number of studies validating Ag-RDTs from different manufacturers is increasing rapidly [1][2][3][4][5][6][7]. However, none have interrogated the performance of Ag-RDTs under conditions that differ from supplier-recommended storage and operation conditions (2− 30 • C), such as those observed in tropical settings where ambient temperatures routinely exceed 30 • C (Fig. 1A). This is challenging because tropical regions are strongly affected by the SARS-CoV-2 pandemic as evident from total cases reported from India, Brazil, Argentina, and Colombia, four out of the ten most affected countries worldwide by November 2020 (Fig. 1B).
On the other hand, the global North was heavily affected by the second wave of the COVID-19 pandemic during November 2020-February 2021 [10,11]. To manage testing demand, different actors have opened external testing stations such as diagnostic streets or drive-through facilities in urban settings [12]. These facilities are often of provisional nature, for example in the form of unheated tents. In the winter months, temperatures in Europe or the U.S. can range from -10 • C to 10 • C [13,14], well below the recommended operating temperatures of most Ag-RDTs. Most manufacturers of SARS-CoV-2 Ag-RDTs specify storage conditions between 2− 30 • C, but stipulate that tests be equilibrated to room temperature (15− 30 • C) at the time of use to guarantee performance. With temperatures around freezing point during the winter months, unheated testing facilities cannot always comply with these conditions. Temperature tolerance of SARS-CoV-2 diagnostic tools or environmental stability requirements have been previously discussed as hurdles to be addressed according to the World Health Organization (WHO) [15,16].
To validate the performance of SARS-CoV-2 Ag-RDTs in both, tropical and cold settings, we compared analytical sensitivity and specificity using recommended conditions and either elevated or low temperatures.

Analytical sensitivity
SARS-CoV-2 (BetaCoV/Munich/ChVir984/2020) was grown on Vero E6 cells (C1008; African green monkey kidney cells), maintained in DMEM (10 % FCS) at 37 • C with 5% CO2. For quantification, viral RNA was extracted using the QIAamp Viral RNA Mini Kit (Qiagen, Hilden, Germany) and quantified using photometrically quantified in vitrotranscribed RNA standards [17,18]. For determining the limit of detection (LOD), SARS-CoV-2 stock (2.2×10 9 copies/mL) was serially diluted in plain DMEM and 5 μL per dilution were added to the extraction buffer of the respective kit for validation experiments. For Coris COVID-19 Ag Respi-Strip, 5 μL of SARS-CoV-2 supernatant were added to 95 μL of PBS to reach the required sample volume of 100 μL prior to addition of LY-S buffer. Validation experiments were performed in triplicates for a subset of tests at recommended conditions ( Fig. 2, setting (i)) initially, with all three replicates showing the same result (Supplementary Table S1). Consequently, due to a limited number of available tests, experiments were performed in duplicates. LOD was defined as the lowest dilution at which both replicates were positive. A dilution factor correction was applied based on the volume of extraction buffer (range: 100− 500 μl) provided by each SARS-CoV-2 Ag-RDT kit.
Specificity for cold conditions was tested using cell culture supernatant of common respiratory viruses including HCoV-229E, HCoV-OC43, influenza virus A H1N1 (7.8×10 6 copies/mL) and rhinovirus A (2.2×10 6 copies/mL). 20 μL of viral cell culture supernatant were added to proprietary lysis buffer or as an internal control 20 μL of lysis buffer were directly applied to test cassettes for validation experiments. Viral concentrations were selected according to the guidelines on analytical specificity testing for SARS-CoV-2 Ag-RDTs published by the German Federal institute for vaccines and biomedicines [19].

Healthy SARS-CoV-2 negative subjects
Additionally, ten healthy laboratory members who previously volunteered for a SARS-CoV-2 Ag-RDT validation study were tested [1]. Healthy volunteers were without symptoms of respiratory tract infection and tested negative for SARS-CoV-2 by RT-qPCR [20]. All subjects received instructions on self-sampling, recently shown to be a reliable alternative to nasopharyngeal swabs taken by professional healthcare workers for Ag-RDTs [21]. Swabs were dissolved immediately in 1 mL PBS and 20 μL of PBS containing respiratory material from study participants were added to proprietary buffer for testing.

Interpretation of test results
For tests with visual readout, results in the form of a band were scored by two researchers independently and in case of discrepancy a third person was consulted to reach a final decision (reader-based tests: Bioeasy 2019-nCoV Ag and ichroma -COVID-19 Ag). Results were defined as borderline when a weak, discontinuous band or smear was observed that could not be clearly defined as a positive or negative result.

World heat map
Data of maximum temperatures of the warmest month ( • C) on country level at the spatial resolution of 2.5 min were obtained from WorldClim 2 [8]. R package 'exactextractr' version 4.0.2 was used to calculate national means. Data on COVID-19 cases were obtained from Worldmeter [9] and visualized using the GraphPad Prism software version 9.1.0.

Results
At present, there are at least 139 SARS-CoV-2 Ag-RDTs commercially available [22], from which 11 were selected for temperature stability validation at elevated temperatures based on the availability of clinical performance data [1] and manufacturing by leading suppliers implying availability on the global market (Table 1).
Subsequently, analytical performance of the selected SARS-CoV-2 Ag-RDTs was assessed following storage and application of tests under recommended conditions as well as elevated temperatures (termed tropical conditions henceforth), using six different experimental settings ( Fig. 2A). The tested conditions were defined by different combinations of storage time (short-and long-term storage) at either recommended (15-30 • C) or elevated temperatures (37 • C) and subsequent test operation at either recommended (15-30 • C) or elevated temperatures (37 • C).
First, we determined analytical sensitivity at recommended conditions by determining the limit of detection (LOD) of SARS-CoV-2 Ag-RDTs when stored and operated at room temperature (15-30 • C; setting (i); Fig. 2) . The dilution-factor corrected LODs for validated SARS-CoV-2 Ag-RDTs ranged from 1.0×10 6 copies/mL to 5.5×10 7 copies/mL of SARS-CoV-2 cell culture supernatant (Table 2). Even though we used a relatively low number of replicates, those LODs were consistent with previously published virus concentrations for validation of SARS-CoV-2 Ag-RDTs using clinical samples [1], suggesting robustness of our data. Our data also highlight profound differences in analytical sensitivity of up to 50-fold for SARS-CoV-2 Ag-RDTs from different manufacturers.
We then assessed analytical sensitivity of SARS-CoV-2 Ag-RDTs after long-term storage at recommended conditions (15-30 • C; 1-6 months) followed by short-term exposure to 37 • C (10 min) and test operation at either 37 • C (setting (ii); Fig   evaluated SARS-CoV-2 Ag-RDTs (five out of eleven; 45 %) was already compromised by about ten-fold when tests were stored under recommended conditions but exposed to 37 • C for only ten minutes prior to testing at 37 • C (condition ii; Fig. 3; for LOD refer to Supplementary  Table S2). This effect was even more pronounced when tests were stored under recommended conditions but exposed to 37 • C for ten minutes prior to testing at recommended temperatures (15-30 • C) (condition iii; Fig. 3), as all eight tested kits (three other kits were not available in sufficient numbers for testing this condition) showed an about 10-fold reduced sensitivity under this experimental setting. We also assessed test performance when Ag-RDTs were stored at recommended temperatures (15-30 • C) followed by direct operation at 37 • C (setting (iv); Fig. 2A). Even under these conditions we found an about ten-fold reduction in analytical sensitivity for six out of eight tested kits (75%; three other kits were not available in sufficient numbers for testing this condition), underlining the importance of test operation at recommended conditions (condition iv; Fig. 3). We additionally tested analytical sensitivity of SARS-CoV-2 Ag-RTDs after long-term storage at 37 • C (3 weeks) followed by test operation at either 37 • C (setting (v); Fig. 2A) or at recommended temperatures (15-30 • C; setting (vi); Fig. 2A). After 19-21 days of storage at 37 • C and testing at 37 • C (condition v, Fig. 3) or testing at recommended temperatures (15-30 • C) (condition vi, Fig. 3), eight out of the total eleven SARS-CoV-2 Ag-RDTs (73 %) showed an about ten-fold reduction in analytical sensitivity when compared to recommended temperatures. In sum, those data indicate that even short-term exposure of SARS-CoV-2 Ag-RDTs to elevated temperatures affects their sensitivity and that multiple temperature shifts might more seriously affect test sensitivity.  Additionally, analytical specificity of SARS-CoV-2 Ag-RDTs under recommended storage and operating conditions and under tropical storage and operating conditions at 37 • C (settings (i) and (v); Fig. 2A) was examined by testing for cross-reactivity with the ubiquitous human coronaviruses HCoV-229E and HCoV-OC43 [23,24]. SARS-CoV-2 Ag-RDTs showed no cross-reactivity with HCoV-229E or HCoV-OC43 upon storage and testing at elevated temperatures ( Table 3).
As the national COVID-19 reference laboratory in Germany, we have been contacted by multiple outside testing facilities across Germany reporting an unusually high number of positive SARS-CoV-2 Ag-RDTs. In order to validate the specificity of SARS-CoV-2 Ag-RDTs when operated at low outside temperatures, we compared Ag-RDT performance after recommended storage and short-term exposure (30 min) to 2− 4 • C prior to testing at 2− 4 • C (setting (vii); Fig. 2B) as well as after storage at 2− 4 • C followed by test operation at 2− 4 • C (setting (viii); Fig. 2B) to Ag-RDT specificity when stored and operated under recommended conditions (setting (i) ; Fig. 2B). We hereto selected a subset of six Ag-RDTs for reasons of scarcity of tests and urgency to conduct the testing under the weather conditions that prevailed at the time of physicians' reports from external testing stations ( Table 1). Two of the six SARS-CoV-2 Ag-RDTs showed impaired specificity (Fig. 4A) when stored at room temperature, but when exposed to 2− 4 • C for 30 min prior to testing at 2− 4 • C (condition vii; Fig. 4A) as cross-reactivity with common respiratory viruses, and false-positive results occurred in healthy volunteers in the form of weak, but clearly visible bands (Fig. 4B). In one test (test I), unspecific reactivity was only observed upon short-term exposure to 2− 4 • C followed by test operation at 2− 4 • C (condition vii; Fig. 4A), but not after long-term storage at 2− 4 • C (condition viii; Fig. 4A). In contrast, the other test which showed non-specific results (test II) yielded almost identically unspecific results after both, short-and long-term storage at 2− 4 • C and test operation at 2− 4 • C (conditions vii and viii; Fig. 4A). On the one hand, those data highlight differences between test devices. On the other hand, our results may hint at effects of relatively rapid temperature changes on some tests for unknown reasons, potentially including environmental factors such as condensation. Results were reproducible and functionality of tests was confirmed by determining their LODs using serial dilutions of SARS-CoV-2 nucleoprotein (SARS-CoV-2-N) at recommended conditions as previously described [1].

Discussion
Our study highlights that temperature during storage and operation of SARS-CoV-2 Ag-RDTs affects test performance. First, even short-term exposure to elevated temperatures may compromise sensitivity of currently available SARS-CoV-2 Ag-RDTs. Our data are consistent with impaired sensitivity of other Ag-RDTs at elevated temperatures, including Malaria rapid diagnostic tests (MRDTs). An assessment of five MRDTs reported a 13%-53% decline in sensitivity for three of those MRDTs following 90 days of storage at 35 • C [25]. Moreover, an evaluation assessing temperature stability of dengue NS1 antigen-based RDTs at 35 • C showed a gradual decline in test sensitivity for seven out of eight tested dengue Ag-RTDs after storage for about 20 days at elevated temperatures [26]. Beyond storage, elevated temperatures during shipment can also affect Ag-RDT performance. Supply chains of MRDTs were studied in Burkina Faso, Senegal, Ethiopia, the Philippines and Cambodia, demonstrating regular exceeding of 30 • C during transport [27,28]. Consequently, the WHO recommends heat stability testing between 35 • C and 40 • C for MRDTs [29] and supply and delivery chains to tropical countries must contain adequate cold chains [30]. Our data thus imply a huge challenge to tropical countries with regard to adequate transportation and storage of SARS-CoV-2 Ag-RDTs. To guarantee temperature-regulated storage, a certain level of infrastructure is required, ideally including air-conditioned facilities with temperature monitors and secured power supply. Unfortunately, these requirements are not always realistic in resource-limited settings and appropriate concepts for adequate storage in remote areas without electricity and rudimentary infrastructure will be required.
Moreover, our study highlights that specificity of SARS-CoV-2 Ag-RDTs may be impaired when operating tests at temperatures below commonly recommended conditions, leading to false-positive results. These results were observed for certain test brands only, including one of the Ag-RDT currently listed for emergency use by the WHO [31,32], highlighting that each test may need to be considered specifically and broader validation of temperature robustness of SARS-CoV-2 Ag-RDTs should be performed. Of note, all tests studied here were shown to be highly specific when operated at recommended conditions in prior studies [1], underlining that impaired specificity is not a test-intrinsic problem but owed to test operation under conditions beyond those defined by the manufacturers. Our data imply that caution must be taken when offering SARS-CoV-2 Ag-RDT-based detection in settings lacking temperature control, including diagnostic streets, drive-through testing stations and self-use by untrained individuals at home [12]. Irrespective of the setting, compliance with the conditions recommended by the manufacturer are vital to ensure accurate testing [33]. As discussed by others, temperature stability guidelines for in vitro diagnostics exist, however there are currently no specific guidelines for the validation of Ag-RDTs regarding temperature stability [34][35][36][37]. Common validation guidelines including environmental conditions could be a first step towards globally reliable diagnostics.
Our study is limited by focusing on analytical test performance for reasons of comparability of test results across the different conditions and based on limited access to clinical samples. An additional limitation of our study is the use of duplicates for some tests instead of a higher number of replicates, which was due to the limited availability of all tests included in the study. Further studies will be required to assess test performance upon storage and application in tropical as well as cold conditions using large numbers of clinical samples. Despite these limitations, our study presents a robust resource for further validation studies as a high number of SARS-CoV-2 Ag-RDTs was included. Additionally, our data on an overall impaired performance of Ag-RDTs at elevated temperatures are consistent across tests and analytical sensitivity for several tests was identical upon usage of either duplicates or higher numbers of replicates.
In sum, it was previously shown that virus concentrations of about 10 6 genome copies per mL suffice for virus isolation and therefore serve as a correlate for infectivity [38,39]. Our study strongly suggests that short-and long-term exposure to elevated temperatures may compromise sensitivity of SARS-CoV-2 Ag-RDTs to an extent that may lead to false-negative test results at clinically relevant virus concentrations, Table 3 Analytical specificity of SARS-CoV-2 antigen-detecting rapid diagnostic tests at recommended and tropical storage and test operation conditions.

Condition
Virus Cps/mL  I  II  III  IV  V  VI  VII  VIII  IX  X  XI i potentially enhancing SARS-CoV-2 spread in tropical settings. At the same time, false-positive test results owed to test operation at low temperatures might not only lead to unwarranted individual quarantine assignments, but also to potential regional lockdown measures if those results were reported to public health authorities without confirmation by a gold standard test such as RT-PCR [20].

Declaration of Competing Interest
The authors report no declarations of competing interest.