The unprecedented spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) resulted in the urgent need for rapid and reliable diagnostic tests. Accurately diagnosing individuals with infection was paramount to limit the transmission of the virus and to reduce morbidity and mortality. Whilst individuals exposed to SARS-CoV-2 appear to be equally at risk of acquiring infection, the severity of the resulting clinical disease differs markedly by age with a case fatality rate of <1% for people <60 years and sequentially increasing to 14.8% among those 80 years or older [1].

Based on the first available SARS-CoV-2 viral sequences, the World Health Organization (WHO) issued guidance on polymerase chain reaction (PCR)-based assays to be performed from upper respiratory tract specimens as the “gold standard” for the detection of SARS-CoV-2 infection [2]. Africa, and in particular, South Africa, relied on the existing PCR-based platforms that had been established for human immunodeficiency virus (HIV) and tuberculosis (TB), enabling the rapid introduction and scale-up of testing for SARS-CoV-2 infections [3]. Notwithstanding South Africa’s diagnostic capabilities to undertake testing, commercially available diagnostic tests and consumables including competitive first-world pricing and prioritisation to specific institutions were a major challenge in accessing and scaling up testing services to address the rapidly growing needs of the country to determine the extent of current and past infection.

The evolution of the rRT-PCR is based on primers and probes (nCoV_IP2 and nCoV_IP4) that were designed to target the genes that encode for the nucleocapsid (N), envelope (E), spike (S), and RdRp proteins [4]. Rapid diagnostic tests (RDT) were designed and to be point-of-care (POC) tests for the detection of SARS-CoV-2 gene/s and/or antigens that are simpler to perform and have a shorter turnaround time. There is minimal evolution of the N gene and therefore most POC tests target the nucleocapsid. Once the virus has entered the host cell, it releases its genomic mRNA material in the cytoplasmic compartment and the translation of ORF-1a and ORF-1b begins [5]. This is followed by viral RNA expression and the replication of genomic RNA to produce full-length copies that are incorporated into newly produced viral particles [6]. Individuals with SARS-CoV-2 infection elicit an innate immune response within hours of viral exposure, followed by the development of Immunoglobulin M (IgM) and Immunoglobulin G (IgG) antibodies at around 7 to 14 days ^[2][7][2,7]. Thus, the detection of SARS-CoV-2 gene/s and/or antigen and antibody responses helps in understanding the infectiousness, transmission dynamics, and natural history of the disease.

Viral shedding has been found to occur in oropharyngeal and nasal or sputum, tracheal aspirates, bronchoalveolar lavage and saliva [8], faeces, urine [9], and semen samples [10]. These findings highlight the need for alternative sampling approaches to improve diagnostic performance and to understand the magnitude and/or duration of viral shedding that could correlate with disease severity and viral dynamics to influence infection and transmission outcomes. The rapidly evolving SARS-CoV-2 pandemic with the emergence of new SARS-CoV-2 variants and subvariants has led to complex diagnostic testing challenges, especially in settings with limited access to diagnostic tests.

2. SARS-CoV-2 Gene and Antigen-Based Diagnostic Tests

Studies that evaluated diagnostic tests for the detection of SARS-CoV-2 genes and antigens included RT-PCR variant genotyping ^[11][14], followed by the Allplex^TM SARS-CoV-2 Variants II multiplex real-time PCR genotyping assay by Seegene (Seoul, South Korea). The testing was based on circulating Beta and Delta variants prior to the emergence of the Omicron variant and utilised specific primers and probes for each variant. The results were available in two hours as opposed to the time-consuming next-generation sequencing. This assay delineated the Beta and Delta variants and had the ability to determine the rapid rate at which the Delta displaced the Beta variant in the study setting of Limpopo, and thus the capability of the assay to rapidly monitor circulating variants ^[11][14]. The reproducibility of the assay was identical across operators with near identical cycle threshold (Ct) values, whilst the overall average Pearson correlation for linearity between the SARS-CoV-2 median Ct and variant typing Ct values for the samples analysed was 0.976 (standard deviation (SD) ±0.019) with 96.4% concordance for repeatability. However, testing was restricted to known circulating variants. To improve the turnaround time and to be less reliant on reagents, equipment, and staff, Marais et al. ^[12][15] from the Western Cape applied a revised workflow using rapid sample preparation (RSP) with a key modification that included sample centrifugation and heating prior to RT-PCR for either the Abbott RealTime SARS-CoV-2 assay or the Allplex^TM 2019-nCoV assay platforms. This modification showed a 97.37% (95% confidence interval (CI):92.55–99.28) positive per cent agreement (PPA) and a 97.30% (95% CI:90.67–99.52) negative per cent agreement (NPA) compared to nucleic acid purification-based testing. In confirmed Delta variant infections, the PPA of RT-PCR on saliva was 73% (95% CI:53.0–84.0). In Omicron variant infections, saliva performed as well as or better than mid-turbinate samples up to day 5, with an overall PPA of saliva swabs of 96% and mid-turbinate samples of 93%, demonstrating the altered kinetics in viral shedding ^[13][16]. As the demand for diagnostic testing overwhelmed the capacity to deliver, Omar et al. assessed the utility of a mobile laboratory staffed with non-laboratory healthcare personnel to undertake PCR testing ^[14][17]. Using the 2400 SARS-CoV-2 Smartchecker PCR kit (Genesystem, Daejeon, South Korea) targeting the N and RdRp genes and processed using the thermocycler (Genechecker; Genesystem, Daejeon, South Korea) showed a median turnaround time of 152 min (interquartile range 123–184) with sensitivity and specificity of 95% and 97% and positive and negative predictive values of 82.4% and 99.2%, respectively, when compared to a clinical diagnosis of COVID-19. With increasing demands on testing for SARS-CoV-2 infection, two studies ^[15][16][18,19], evaluated the field performance of the Abbott Panbio Antigen Rapid Test Device (Ag-RDT) (Abbott, San Diego, Carlsbad, CA, USA) against the available SARS-CoV-2 RT-PCR, which detects the Beta and Delta variants ^[15][16][18,19]. In the Eastern Cape province, the test had a sensitivity of 69.17% (95% CI:61.4–75.8) and specificity of 99.02% (95% CI:98.8–99.3) among symptomatic individuals ^[15][18], whilst among members of the public at three taxi ranks in Johannesburg, the test had a sensitivity of 40.0% (95% CI:30.3–50.3) and specificity of 98.5% (95% CI:96.9–99.4) with a positive predictive value of 85.1% (95% CI:71.7–93.8) and a negative predictive value of 88.5% (95% CI:85.5–91.1) ^[16][19]. The sensitivity of the test was dependent on the amount of viral RNA in clinical samples, as reflected by the PCR Ct value ^[16][19].

3. SARS-CoV-2 Antibody-Based Diagnostic Tests

Serological assays for the detection of IgG, IgM, or Immunoglobulin A (IgA) against SARS-CoV-2 infection provide important information for surveillance, antibody persistence, infection rate, and vaccine coverage. Serological assays, including enzyme-linked immunosorbent assays (ELISA) and rapid lateral flow assays, are available commercially; however, the high cost limits their accessibility in resource-limited countries. Although several assays have been developed, field evaluations have been limited. Testing was performed on serum samples, plasma, fingerstick, and dried blood spot (DBS) samples. Of the twelve studies, five (38.5%) compared serological outcomes to RT-PCR, whilst in seven (54%), a comparison was made to either in-house or commercially available serological tests. Makatsa et al. (2021) ^[17][20] developed an in-house indirect ELISA using plant-derived recombinant viral proteins by means of the S1 and receptor-binding domain (RBD) portions of the spike protein from SARS-CoV-2, expressed in Nicotiana benthamiana ^[17][20]. This test measured antibody responses among SARS-CoV-2 PCR-positive patients. Samples taken at a median of 6 weeks from diagnosis from patients with mild and moderate COVID-19 disease showed that the in-house ELISA, when compared to the S1 IgG ELISA kit (EUROIMMUN), detected immunoglobulins; S1-specific IgG was detected in 66.2% and RBD-specific IgG in 62.3% of samples and were concordant with the EUROIMMUN assay. To optimise the diagnostic algorithm for SARS-CoV-2 infection, Gededzha et al. (2021) ^[18][21] evaluated the diagnostic performance of the EUROIMMUN Anti-SARS-CoV-2 ELISA for the semi-quantitative detection of IgA and IgG antibodies in serum and plasma samples targeting the recombinant spike (S1) domain of the SARS-CoV-2 spike protein as the antigen. The sensitivity of EUROIMMUN was higher for IgA (74.3%, 95% CI:69.6–78.6) than for IgG (64.1%, 95% CI:59.1–69.0), though specificity was lower for IgA (84.2%, 95% CI:77–89.2) than IgG (95.2%, 95% CI:90.8–98.4) and both sensitivity and specificity improved in symptomatic individuals ^[18][21]. The performance of the Abbott SARS-CoV2 Architect and Abbott SARS-CoV2 Alinity IgG when compared to RT-qPCR showed the sensitivity of the assays to be 69.5% (95% CI:64.7–74.1) and 64.8% (95% CI:59.4–69.9), respectively, whilst the specificity of the assays was 95% (95% CI:89.9–98) and 90.3% (95% CI:82.9–95.2), respectively. When the assays were compared to the in-house ELISA, the sensitivity for the Architect and Alinity assays was 94.7% (95% CI:88.8–98) and 92.5% (95% CI:85.8–96.7), respectively, whilst specificity was 88.1% (95% CI:79.2–94.1) and 91.7% (95% CI:83.6–96.6), respectively. The sensitivity for both assays was highest at 31–40 days post-presentation and lowest at time points of less than 7 days. These findings highlight the futility of testing for antibody responses during the acute and early stages; that is, within less than 14 days of infection ^[19][22]. Matefo et al. (2022) investigated two in-house ELISAs and an in-house immunofluorescent assay (IFA), developed using the SARS-CoV-2 S1 protein, for use in South African populations ^[20][23]. The tests were compared with Roche Elecsys^TM Anti-SARS-CoV-2 (Roche Diagnostics GmbH, Mannheim, Germany) and a commercial lateral flow assay, COVID-19 IgG/IgM Rapid Test cassette (Zhejiang Orient Gene Biotech Co., Ltd., Zhejiang, China). Based on IgG antibodies, specificity was 96% and 100% for ELISA and IFA, respectively, and sensitivity was shown to be 100% and 98.8% for ELISA and IFA, respectively, for samples collected one week after the onset of illness. Positive predictive values were 92.1% for ELISA and 91.0% for IFA. The in-house ELISA and IFA were positive for IgG antibodies, regardless of circulating variants, therefore demonstrating the potential of these tests for high throughput screening in resource-constrained environments ^[20][23]. The performance of the Roche Elecsys^TM chemiluminescent immunoassay (Rotkreuz, Switzerland) to detect antibodies to SARS-CoV-2 N as antigen was evaluated by Grove et al. ^[21][24]. Among patients from Johannesburg, serum samples from SARS-CoV-2 RT-PCR positive and negative individuals showed a sensitivity of 65.2% (95% CI:59.57–70.46) and specificity of 100% (95% CI:97.07–100). The sensitivity of the test improved to 72% among those with >14 days and to 88.6% in those 31–50 days post diagnosis. Nevertheless, using the in-house ELISA assay utilising the plant-based S1 and RBD as antigens ^[17][20], the overall PPA was 89.4% (95% CI:82.18–94.39) and NPA was 88.4% (95% CI: 80.53–93.83). However, among individuals at earlier time points post-infection and among asymptomatic individuals, the sensitivity was lower with the Roche Elecsys^TM chemiluminescent immunoassay and the in-house ELISA ^[21][24]. David et al. (2021) evaluated 30 lateral flow immunoassays using serum or plasma samples from patients with confirmed SARS-CoV-2 infection ^[22][25]. Of these, 26 assays did not meet the predefined operational acceptance criteria for kits to be approved for use in South Africa. Whilst the performance of the lateral flow tests was similar to the sensitivities and specificities reported in other studies, only four (13%) assays (Zheihang Orient Gene COVID-19 IgG/IgM, Genrui Novel Coronavirus (2019-nCoV) IgG/IgM, Biosynex COVID-19 BSS IgG/IgM, Boson Biotech 2019-nCoV IgG/IgM) were recommended for South Africa Health Products Regulatory Authority (SAHPRA) approval ^[22][25]. Among volunteers in Cape Town, 23.7% tested positive for IgG antibodies with the Abbott SARS-CoV-2 IgG assay. Of those who tested positive, 47.9% reported no symptoms of COVID-19 in the past 6 months. Seropositivity was significantly associated with living in informal housing, residing in a subdistrict with low income per household, and having a low-earning occupation. The specificity of the assay was 98.54% (95% CI:94.82–99.82) ^[23][26]. In the household survey undertaken in three communities across three provinces in South Africa, the burden of SARS-CoV-2 infections was measured using two ELISA kits: Wantai SARS-CoV-2 Ab ELISA (Beijing Wantai Biological Pharmacy Enterprise), measuring total antibodies (IgM, IgG and IgA) against the RBD in the spike protein, and Roche Elecsys^TM Anti-SARS-CoV-2 ELISA (Roche Diagnostics), measuring total antibodies to the N protein. There was 94.5% PPA with a Cohen κ statistic of 0.89. The Wantai assay, compared with the Roche Elecsys^TM assay, had a sensitivity of 91.0% and a specificity of 97.2% ^[24][27] To monitor antibody responses to SARS-CoV-2 following a vaccine rollout, Maritz et al. (2021) assessed a ligand binding-based serological assay for the semiquantitative detection of IgG, IgM, IgA, and neutralising antibodies (nAb) in serum ^[25][28]. The assay demonstrated high levels of diagnostic specificity and sensitivity (85–99% for all analytes). Serum IgG, IgM, IgA, and nAb correlated positively (R2 = 0.937, R2 = 0.839, R2 = 0.939 and R2 = 0.501, p < 0.001, respectively) with those measured in DBS samples. In vitro SARS-CoV-2 pseudotype neutralisation correlated positively with the solid phase nAb signals in convalescent donors (R2 = 0.458, p < 0.05), highlighting the potential use of the assay in efficacy studies, infection monitoring, and post-marketing surveillance following vaccine rollout ^[25][28]. To enable large-scale testing for SARS-CoV-2 antibodies, DBS samples were evaluated against plasma samples with a correlation of r = 0.935 and 0.965 for RBD and full-length S-protein of SARS-CoV-2 ^[26][29]. A Bland–Altman assessment showed agreement between IgG mean fluorescence intensity (MFI) values with 6.25% of observations for both RBD IgG and spike IgG, falling outside the 95% limit of agreement. Therefore, DBS samples are a useful medium for population screening and field studies in resource-constrained settings, as they are non-invasive and ideal for storage, transportation, and processing ^[26][29]. As the need for testing increases, especially for surveillance or during outbreak situations, rapid antibody testing assays are useful in such situations. Irwin et al. (2021) evaluated the sensitivity of five rapid antibody assays and explored factors influencing their sensitivity in detecting SARS-CoV-2-specific IgG and IgM antibodies ^[27][30]. In addition, finger-prick blood samples from participants within 2–6 weeks of PCR-confirmed COVID-19 diagnosis were included in the evaluation. Overall sensitivity for IgG and IgM antibodies was below 70% and ranged from 13% to 67% for IgG and markedly lower for IgM. Whilst rapid tests in resource-constrained settings are a promising tool in COVID-19 diagnosis, the sensitivity was reduced for those under 40 compared with those over 40 years of age. These findings show significant variability when used in real-world settings, limiting their application ^[27][30].