This case may be the first patient to contract COVID-19 in California, and perhaps the third in the United States, given that the third and fourth cases in the country were diagnosed on January 26 in California [4]. Serology testing may have had utility in acute diagnosis, particularly in patients with atypical presentations such as this case. In addition, there is a rapidly growing need for serology testing as many patients appear to contract COVID-19 infection without becoming symptomatic. Guidelines remain unclear as to when to order a qualitative RDT serology test and how to interpret the results, therefore we propose the following recommended considerations and approach.
Pre-test probabilities should be factored into laboratory testing decisions and should appropriately influence physician interpretation of results [5]. The very low U.S. prevalence of COVID-19 in late January, combined with an atypical presentation (78% of Chinese patients describe fever as the primary symptom, followed by persistent cough) [6], and lack of risk factors such as travel to China, may have led to a result more likely to be a false positive than a true positive. The hsCRP and D-dimer level results could be considered non-specific given their modest specificities [7]. It should be noted, however, that D-dimer levels > 2.0 µg/mL have been reported as a strong predictor of COVID-19 in-hospital mortality, possibly reflecting activation of coagulation and fibrinolysis [8]. A positive finding of influenza hemagglutination inhibition (HAI) antibodies in a previously immunized person would not differentiate between acute influenza infection and HAI from the vaccination, and a negative result would not necessarily increase suspicion for COVID-19, given the limited 45% estimated efficacy rate of the influenza vaccine in 2019-2020 [9]. The CT scan finding of unilateral vs. bilateral infiltrate and effusion is another somewhat atypical finding, although approximately 20% of Chinese COVID-19 patients present with unilateral rather than bilateral radiological findings [10,11]. However, as the pre-test probability of COVID-19 infection has currently risen to levels many times higher than expected in the medical community, practitioners today should have a commensurately high index of suspicion for COVID-19 in the differential diagnosis. Two studies in April provide striking illustrations: first, studies a PCR-based screening of 3,300 patients in Santa Clara County, California found a demographic and test sensitivity adjusted prevalence of 4.2% [12], and second, a New York-based study reported 13.7% community prevalence based on universal screening of 215 pregnant women [13].
As the community prevalence of COVID-19 explodes, the question as to whether asymptomatic or mildly symptomatic healthcare professionals were previously infected has become more acute. Practitioners live with high states of anxiety related to surviving in a work environment with high levels of exposure to infected patients. Knowledge of prior infection might ease these concerns, with the important caveat that qualitative RDT serology tests do not provide quantitative IgG titers as enzyme linked immunosorbent assays (ELISA) do, and neither test guarantees immunity. To establish immunity, a 3-5 day neutralization assay would have to show that the patient’s antibodies can inhibit viral growth in a cell culture system. However, there are no FDA-approved/authorized neutralization assays for SARS-CoV-2 in the U.S, these assays may return false negative results if they miss antibodies to viral proteins not involved in replication, and a recent Chinese study found poor correlations between high titers of neutralizing antibodies and efficacy retarding virus growth in cell culture [14]. Although IgM-/IgG+ results indicate past infection, it is not established that such patients cannot continue to transmit the virus. Current CDC guidance for healthcare workers after positive COVID-19 diagnosis (which may be based on clinical impression or test result), include a test-based strategy of obtaining two negative PCR tests 24 hours apart to rule out SARS-CoV-2 viral loads in the nasopharynx, or a non-test based strategy in lab-test confirmed COVID-19 positives who are asymptomatic to wait ten days after their test result so long as they do not develop symptoms during that interval [15].
Positive serology results may indicate either current and recent infection (IgM titer elevated) or prior infection (IgM-/IgG+). Because SARS-CoV-2 viral RNA loads rise before IgM levels, PCR tests are preferred in early acute infection, and serology tests may have a significant false negative rate as illustrated in Fig. 3.
After an incubation period averaging 5 days (range 1-2 days to 14 days), COVID-19 symptoms develop [16]. The exact time course for the rise in viral RNA vs. IgM/IgG titers is not yet broadly studied and may vary if different serology assay have different limits of detection. Guo et al. found that 22% of serology tests were negative in symptomatic COVID-19 patients positive by PCR, and that IgM peaked 5 days after symptoms onset vs. IgG at 14 days [17]. The false negative rate for serology testing may also be related to symptom severity, with Wu et al. reporting that 30% of COVID-19 recovered patients never developed high levels of neutralizing IgG antibodies [14]. In the Guo et al. study there was no cross-reactivity against the “benign” coronavirus strains NL63, 229E, OC43 and HKU1 which are responsible for one-third of common colds [18]. Although false positives could be caused by cross-reactivity of SARS-CoV-2 test antibodies to SARS-CoV-1, this is unlikely to cause a false positive in the U.S., given its very low incidence of SARS cases in 2003-2004 [19]. Other serology tests may not achieve the same specificity as the assay cited above, thus clinicians should be alert to potential assay differences in cross-reactivity to common cold coronavirus strains [20].
In order to expedite test availability, The FDA has not insisted on reviewing and authorizing all serology tests until May 4 [21]. Of the many dozens of different tests currently promoted commercially, only three serology tests achieved FDA Emergency Use Authorization (EUA) as of April 15. Importantly, FDA EUAs for both PCR and serology tests do not require reporting of clinical sensitivity and specificity based on COVID-19 samples split and compared to a reference standard. Instead EUAs are issued based on analytic performance using contrived samples in test tubes. This approach has important limitations, for example analytic performance review only for a PCR test may not reveal limitations in clinical sensitivity caused by inadequate sampling of the nasopharynx, and contrived samples used to evaluate a serology test that lack common cold coronavirus antigens may not reveal problems with clinical specificity. Published test performance would not have predicted the likely false negative 2nd serology test in this physician-patient where the lab technician did not wait the recommended 8-10 minutes for the test to turn positive. In general, we expect clinical test performance to be lower than the analytical sensitivity and specificity reported in FDA Instructions for Use for COVID-19 tests.
Clinical index of suspicion for COVID-19 should now be high as prevalence has risen quickly, and as it has been systematically underestimated due to narrow criteria limiting testing to the sickest patients. As many as 44%-79% of cases may be transmitted by asymptomatic patients who would never be selected for PCR testing because of the lack of symptoms of a viral syndrome [22,23]. Many more, like the case here, may present with atypical or mild symptoms and serologic testing could facilitate retrospective diagnosis of COVID-19. PCR vs. serology test selection should be related to duration of symptoms, with the PCR test reserved for acute cases ill for ≤ 7 days, especially in a country where access to PCR testing is limited because of test kits or reagents, inadequate staffing, or supplies of nasal swabs. Antibody testing has been shown to be more sensitive than viral nucleic acid detection after approximately eight days of COVID-19 illness duration in two separate studies [17,24]. Patients with symptoms for more than 7 days might be confirmed with positive IgM titers on a serology test. However, symptomatic IgM-/IgG- patients may be falsely negative, and we would recommend PCR testing in this scenario. Persistence of IgM elevation in either IgM+/IgG- patients or IgM+/IgG+ serology indicates active infection and these patients should be isolated and their contacts traced. IgM-/IgG+ results indicate past infection and isolation is unnecessary. However, IgM-/IgG+ results on a qualitative RDT serology test should not be considered a guarantee of SARS-CoV-2 immunity, and contagion safety precautions should be maintained. In IgM-/IgG+ cases by qualitative RDT testing, a follow-up quantitative IgG ELISA with a high antibody titer might support the argument for immunity, similarly to the way higher rubella antibody titers in pregnant women indicate immunity [25], however this recommendation requires further study. However, this recommendation, like others above, are limited by the general lack of available information on COVID-19, as well as inadequate information on test performance, even for FDA-authorized tests.