Professional Documents
Culture Documents
Mercado, Marcela - 2020
Mercado, Marcela - 2020
of Public Health
Suggested citation Mercado M, Malagón-Rojas J, Delgado G, Rubio VV, Muñoz Galindo L, Parra Barrera EL et al. Evaluation of nine serological
rapid tests for the detection of SARS-CoV-2. Rev Panam Salud Publica. 2020;44:e149 https://doi.org/10.26633/RPSP.2020.149
ABSTRACT Objective. To evaluate the operative capacity of nine serological rapid tests to detect the IgM/IgG antibodies
response in serum from patients with SARS-CoV-2 in different clinical stages.
Methods. A cross-sectional study of serological rapid tests was designed to compare the performance of the
evaluated immunochromatographic tests for the diagnosis of SARS-CoV-2. A total of 293 samples was used,
including negatives, asymptomatic, and symptomatic serum samples.
Results. The sensitivity of the evaluated tests was low and moderate in the groups of asymptomatic serum
samples and the group of serums coming from patients with less than 11 days since the onset of the symp-
toms. The specificity for the anti-SARS-CoV-2 antibodies tests ranged between 86.5%-99% for IgM and
86.5%-99.5% for IgG. The sensitivity and the likelihood ratio were different according to the study groups. The
usefulness of these tests is restricted to symptomatic patients and their sensitivity is greater than 85% after
11 days from the appearance of symptoms.
Conclusions. Serological tests are not an adequate strategy for the identification of asymptomatic and
pre-symptomatic patients. Serological rapid tests for the detection of specific anti-SARS-CoV-2 antibodies can
be used as a diagnostic aid, but diagnosis must be confirmed by RT-PCR. Rapid tests should be reserved for
patients with symptoms lasting more than 11 days.
Since the SARS-CoV-2 atypical pneumonia was reported in cases at the community level with some risk such as hospitals,
December 2019 in China (1), the world has focused on research nursing homes, penitentiary centers, and passenger transport
toward a better understanding of the virus and the infection in airports and terminals (3,4).
in humans, including the development of diagnostic tests for Serological rapid tests are particularly important for early
early detection, treatment, and reducing mortality associated detection; these tests are easy to operate and allow a rapid
with the virus (2). In the current context of the pandemic, one screening of SARS-CoV-2 antibodies within 10-15 minutes
of the main challenges is the early detection of SARS-CoV-2 (5), helping to have a better understanding of the humoral
1
Instituto Nacional de Salud, Bogotá, Colombia. 4
Instituto Distrital de Ciencia, Biotecnología e Innovación en Salud, Bogotá,
2
Facultad de Medicina de la Universidad El Bosque, Bogotá, Colombia. Colombia.
* Jeadran Malagón-Rojas, [email protected] 5
Instituto de Evaluación Tecnológica en Salud, Bogotá, Colombia.
3
Universidad Nacional de Colombia, Bogotá, Colombia.
This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs 3.0 IGO License, which permits use, distribution, and reproduction in any medium, provided the
original work is properly cited. No modifications or commercial use of this article are permitted. In any reproduction of this article there should not be any suggestion that PAHO or this article endorse any specific organization
or products. The use of the PAHO logo is not permitted. This notice should be preserved along with the article’s original URL.
response of infected patients (6). Because of their low cost and According to the manufacturing information, these tests can
easy processing, in some scenarios these types of tests have detect anti-SARS-CoV-2 antibodies in whole venous blood,
been used as a screening tool for the general population. How- whole blood by capillary puncture, serum, and plasma sam-
ever, some manufacturers have stated that the serological rapid ples. According to the manufacturer's instructions, the test is
test requires the concomitant use of RT-PCR due to the false- positive when a red band is generated, one for IgM, another for
negative rate that can result (7). IgG, and one for control (Table 1).
Serological rapid tests generally use a conjugate made up of a
reagent (usually colloidal gold) that is activated in the presence Clinical data
of viral antigens or antibodies against human IgM/IgG. Acti-
vation is visible on the nitrocellulose plate in a colored band. The sociodemographic and clinical data were obtained from
Currently, a large number of serological tests are available, the epidemiological reports from the National Epidemiological
which present different operational capabilities for the detection System (SIVIGILA). It included age, sex, place of residence, type
of SARS-CoV-2 antibodies. However, this detection capacity of infection (asymptomatic and symptomatic), date of onset of
has been questioned by the World Health Organization (8). The symptoms, severity of infection, etc.
aim of this study was to evaluate the operative capacity of sero-
logical rapid tests to detect the IgM/IgG antibodies response in Collection of serum samples
serum of patients with SARS-CoV-2 in different clinical stages.
With the results of molecular tests taken from samples of
METHODS nasopharyngeal swabs in different groups of individuals
(symptomatic, asymptomatic, and negative), a cohort of posi-
A cross-sectional study of the diagnostic tests was designed tive patients was followed up on the 7th, 14th, and 21st day after
to compare the performance of nine immunochromatographic the first nasopharyngeal swab. This was carried out to obtain
tests for the diagnosis of SARS-CoV-2. A minimum sample size blood samples to determine the presence of antibodies over
of 293 was estimated according to the sensitivity parameter 14 days post-infection.
(9), and a margin error of 10%, with a confidence level of 95%.
The available serum was distributed in four groups: i) patients Test procedure
with SARS-CoV-2 and positive RT-PCR who developed mild,
moderate, and severe COVID-19; ii) asymptomatic patients Before measuring and verifying the humidity and tempera-
with positive RT-PCR; iii) patients with negative RT-PCR tests; ture described on the inserts of each test, 10 μL or 5 μL of serum
and iv) “historic negatives”, i.e., serum collected before the was served in the well. Subsequently, 2 or 3 drops of running
pandemic in 2017-2018 and stored in the Instituto Nacional de buffer were added. The presence of movement in the nitrocel-
Salud biobank, which was used as a control for the tests since lulose column was observed after the incubation time of 10 to
these should not have anti-SARS-CoV-2 antibodies (10). 15 minutes, and the reading was then performed according to
the manufacturer's instructions.
Exclusion criteria
Test evaluation
Serum samples from patients were excluded if they had a his-
tory of autoimmune disease such as rheumatoid arthritis, lupus The cassettes results were interpreted and reported by two
erythematosus, and Sjögren syndrome, among others. Like- independent evaluators. The results of the readings were tran-
wise, serum was excluded in those who received transfusions scribed by each of the evaluators independently. Subsequently,
in the last month before taking the blood sample, pregnant a third researcher reviewed the database to check for transcrip-
women, and minors. tion errors or inconsistencies.
As the gold standard technique real-time PCR test with Point and interval estimates of 95% of the operational char-
reverse transcriptase (RT-PCR, Diagnostic detection of 2019- acteristics of the tests concerning sensitivity, specificity, positive
nCoV by real-time RT-PCR, Charité Virology, Berlin, Germany) likelihood ratio, positive and negative predictive value (PPV
was used. RT-PCR consists of a polymerase chain reaction that and NPV respectively) were established (12,13). Also, the group
has previously had a reverse transcription (RT) phase. Using of historical negatives was used to estimate the divergent crite-
RT, cDNA was obtained from an RNA strand. Consequently, ria validity.
the RT-PCR technique performs the detection and amplification A pooled analysis was performed for sensitivity and speci-
of a sequence from a strand of RNA. This presents a technical ficity. Also, I2 was estimated. An I2 >50% was considered as
limit of detection (LOD) = 5.2 RNA copies/reaction, at a 95% moderate heterogeneity; an I2 ≥ 75% was considered high (14,15).
hit rate; 95% CI: 3.7-9.6 copies of RNA/reaction, according to According to the observed heterogeneity (I2), subgroups
the protocol (11). were formed excluding tests involved in the high inconsistency.
Forest plots with test-specific and overall point estimates and
Type of tests used 95% confidence intervals (CI) were carried out. The analysis
was performed with SPSS® V. 16.1 statistical program, licensed
The immunochromatographic assay was used for the qual- by the INS. Forest plots were made in Meta-DiSc® V. 1.4, free
itative detection of SARS-CoV-2 specific IgG/IgM antibodies. license (16).
Ethical considerations between 86.5% and 99% for IgM, and 86.5% and 99.5% for IgG.
However, the sensitivity and the likelihood ratio varied accord-
The study was approved by the Ethics Committee of the INS ing to the study groups.
in June of 2020. Regarding the measurement of the validity of the divergent
criteria (VDC), when evaluated with the historical negative
RESULTS serums, VDC presented values greater than 95% for IgM and
96% for IgG in 8 of the 9 tests evaluated. In the Leccurate® test
A total of 293 serum samples were included in the study, this value was 84% (95%
CI 75.4 - 91.4) for IgM and 86% (95% CI
including 100 historical negatives. In the 2020 group, most of 77.7 - 91.6) for IgG, which should be analyzed in light of the pos-
the samples corresponded to male patients (65.8%, n=127), with sibility of a large number of false positives that may be related to
an average age of 38 years (SD 12). 48.2% (n=93) out of the sam- cross-reaction to other type of antibodies non-specific to SARS-
ples corresponded to positive serums (table 2). CoV-2 (table 3).
Nine immunochromatographic assays were evaluated.
Table 3 shows the performance results of each test in three Asymptomatic patients
groups of serum: asymptomatic patients, patients with ≤ 11
days of symptoms, and patients with ≥ 11 days of evolution. In the nine tests evaluated, the sensitivity of the IgM and IgG
In general, the tests showed adequate specificity to rule out in samples from asymptomatic patients was less than 50%. In
the presence of IgM and IgG antibodies when anti-SARS-CoV-2 contrast, specificity was greater than 85%. The likelihood ratio
immunoglobulins were not present. The specificity ranged had values between 3.50 and 33.
Test Type of Place of Type of test Reading time n Sensitivity Specificity Sensitivity Specificity
cassette manufacturing (min) IgM (%) IgM (%) IgG (%) IgG (%)
Biosensor® Duo Republic of Korea Colloidal gold 10-15 40 91.70 100 79.20 100
AMS International® One United Kingdom Colloidal gold 10-15 360 95.70 97.30 91.80 96.40
Leccurate® One China Colloidal gold 10-20 NR 90 100 NR NR
HIGHTOP One Step One China Colloidal gold 10-20 1300 82 96 93 97.50
rapid test®
Cromatest COVID-19® One China Colloidal gold 10-15 551 78.03 99.39 96.86 100
AMP Rapid Test® One Austria Colloidal gold 10 NR 95.70 97.30 91.80 96.40
Egens® One China Colloidal gold 10 NR 96.88 100 96.88 100
Cellex® One United States Colloidal gold 15-20 378 93.75 96.40 93.75 96.40
Onesite Rapidtest® One United States Colloidal gold 10-15 551 78.03 99.39 96.86 100
NR, not reported
(Continued)
Symptomatic patients between 8-11 days or less The pooled specificity was 0.97 (CI 95% 0.96 – 0.98); p=0.10;
from the onset of symptoms I2=40.6% (Figure 1). Excluding the results from Leccurate, the
pooled sensitivity for IgG was 0.17 (CI 95% 0.12 – 0.21) p=0.99;
The sensitivity was low, between 0%-64.2% for IgM and I2=0%; the pooled specificity was 0.97 (CI 95% 0.96 – 0.98);
between 11.11%-33.3% for IgG. The likelihood ratio did not p=0.12; I2=38.6% (Figure 2).
exceed 22 for IgM and 66.66 for IgG. Symptomatic patients between 8-11 days or less from the onset
of symptoms. The pooled sensitivity for IgM in this group
Symptomatic patients with more than 11 days since was 0.33 (CI 95% 0.23 – 0.43); p=0.007; I2=61.7%. The pooled
the onset of symptoms specificity for IgM was 0.95 (CI 95% 0.94 – 0.96) with a high
inconsistency (I2=81.8%). The PPV and NPP were 0.27 (CI 95%
The sensitivity was low to high, between 37.5%-93.75% for 0.21 – 0.35) and 0.96 (CI 95% 0.96 – 0.97) respectively. The pooled
IgM and between 70.83%-93.75% for IgG. Likewise, the positive sensitivity for IgG was 0.25 (CI 95% 0.16 – 0.35); p=0.86; I2=0%.
likelihood ratio did not exceed 45.83 for IgM and 166 for IgG. The pooled specificity was 0.96 (CI 95% 095 – 0.97) with a high
inconsistency (I2=84.7%). The pooled PPV was 0.25 (CI 95%
Pooled analysis 0.18 – 0.34) and the pooled NPV was 0.96 (CI 95% 0.95 – 0.96).
When we ran the analysis excluding Leccurate, the pooled
A pooled analysis was performed in the asymptomatic and sensitivity for IgM was 0.30 (CI 95% 0.28 – 0.41); p=0.009;
symptomatic groups (less than 11 days and more than 11 days). I2=62.6% (Figure 1). The pooled specificity was 0.97 (CI 95%
Asymptomatic patients. The pooled sensitivity of IgM was 0.95 – 0.98); p=0.10; I2= 40.6%). The pooled specificity for IgG
0.286 (CI 95% 0.238 – 0.337); p=0.002; I2=67.9%. The pooled was 0.25 (CI 95% 0.16 – 0.36); p=0.79; I2= 0%. The grouped
specificity was 0.95 (CI 95% 0.94 – 0.96), with a high heteroge- specificity was 0.97 (CI 95% 0.96 – 0.98); p=0.12; I2=38.6%
neity (I2=81.8%). The pooled PPV was 0.55 (CI 95% 0.48 – 0.61) (Figure 2).
and NPV was 0.87 (CI 95% 0.87 – 0.88). The pooled sensitiv- Symptomatic patients with more than 11 days since the onset
ity of IgG was 0.19 (CI 95% 0.15 – 0.24) p=0.18; I2=29.6%. The of symptoms. The pooled analysis for IgM in this group was
pooled specificity was 0.96, but showed a high heterogeneity 0.66 (CI 95% 0.61 – 0.70) with a high heterogeneity (I2=84.3%).
(I2=84.7%). The pooled PPV and NPV were 0.50 (CI 95% 0.42 – The specificity was 0.95 (CI 95% 0.94 – 0.96), also with a high
0.57) and 0.86 (CI 95% 0.85 – 0.86) respectively. heterogeneity (I2=81.8%). The pooled PPV was 0.78 (CI 95%
When the Leccurate essay was excluded from the analysis 0.74 – 0.82); the pooled NPV was 0.92 (CI 95% 0.91 – 0.93).
there was a reduction in the heterogeneity, and the pooled sen- The pooled sensitivity for IgG was 0.77 (CI 95% 0.73 – 0.81);
sitivity for IgM was 0.26 (CI 95% 0.21 – 0.31); p=0.010; I2 =62.4%. p=0.001; I2=69.6%. The specificity was 0.96 (CI 95% 0.95 – 0.97)
FIGURE 1. Pooled sensitivity and specificity excluding Leccurate of IgM: (A) in asymptomatic group results; (B) in symptomatic
group (< 11 days).
Pooled Sensitivity = 0,26 (0,21 to 0,32) Pooled Specificity = 0,97 (0,96 to 0,98)
Chi-square = 18,60; df = 7 (p = 0,0095) Chi-square = 11,78; df = 7 (p = 0,1082)
Inconsistency (I-square) = 62,4 % Inconsistency (I-square) = 40,6 %
0 0,2 0,4 0,6 0,8 1 0 0,2 0,4 0,6 0,8 1
Sensitivity Specificity
Pooled Sensitivity = 0,30 (0,21 to 0,42) Pooled Specificity = 0,97 (0,96 to 0,98)
Chi-square = 18,70; df = 7 (p = 0,0091) Chi-square = 11,78; df = 7 (p = 0,1082)
Inconsistency (I-square) = 62,6 % Inconsistency (I-square) = 40,6 %
0 0,2 0,4 0,6 0,8 1 0 0,2 0,4 0,6 0,8 1
Sensitivity Specificity
FIGURE 2. Pooled sensitivity and specificity excluding Leccurate of IgG: (A) in asymptomatic group; (B) in symptomatic group
(< 11 days); and (C) in symptomatic group (>11 days).
Pooled Sensitivity = 0,30 (0,21 to 0,42) Pooled Specificity = 0,97 (0,96 to 0,98)
Chi-square = 18,72; df = 7 (p = 0,0091) Chi-square = 11,78; df = 7 (p = 0,1082)
0 0,2 0,4 0,6 0,8 1 Inconsistency (I-square) = 62,6 % 0 0,2 0,4 0,6 0,8 1 Inconsistency (I-square) = 40,6 %
Sensitivity Specificity
Pooled Sensitivity = 0,76 (0,71 to 0,80) Pooled Specificity = 0,98 (0,97 to 0,98)
Chi-square = 16,05; df = 7 (p = 0,0247) Chi-square = 11,39; df = 7 (p = 0,1223)
0 0,2 0,4 0,6 0,8 1 Inconsistency (I-square) = 56,4 % 0 0,2 0,4 0,6 0,8 1 Inconsistency (I-square) = 38,6 %
Sensitivity Specificity
Pooled Sensitivity = 0,76 (0,71 to 0,80) Pooled Specificity = 0,98 (0,97 to 0,98)
Chi-square = 16,05; df = 7 (p = 0,0247) Chi-square = 11,39; df = 7 (p = 0,1223)
0 0,2 0,4 0,6 0,8 1 Inconsistency (I-square) = 56,4 %
0 0,2 0,4 0,6 0,8 1 Inconsistency (I-square) = 38,6 %
Sensitivity
Specificity
with high inconsistency (I2=84.7%). The pooled PPV and NPV in the asymptomatic group (IgM and IgG), were the sensitivity
were 0.83 (CI 95% 0.79 – 0.86) and 0.94 (CI 95% 0.93 – 0.95). was less than 40%.
The pooled sensitivity for IgM excluding the Leccurate These findings are similar to those reported in the literature.
test was 0.62 (CI 95% 0.57 – 0.67); however, the inconsistency The study from Vidal-Anzardo et al (10) comparing serolog-
remained high (I2=74.9%). The pooled specificity was 0.96 ical rapid tests against molecular tests in a population of 143
(CI 95% 0.95 – 0.97); p= 0.10; I2=40.6%. In the case of IgG, patients from Peru reported that the sensitivity of the serologi-
the pooled sensitivity was 0.75 (CI 95% 0.70 – 0.79); p=0.02; cal rapid test is high from the second week of symptoms. On the
I2=56.4%. The pooled specificity was 0.97 (CI 95% 0.96 – 0.98); other hand, it has also been reported better performance of the
p=0.12; I2=38.6% (Figure 2). test in patients who presented clinical symptoms longer than
12 days (17,18). Zhang et al reported that the utility of these sero-
DISCUSSION logical tests is increased in patients with at least 8 days of clinical
evolution (19). The author identified that the use of serological
In this study, we evaluated nine commercial immunochro- tests for the detection of specific anti-SARS-CoV-2 antibodies in
matographic tests for the detection of anti-SARS-CoV-2 IgM febrile patients is useful to carry out differential diagnosis, but
and IgG antibodies. Our results showed that the best sensitiv- that it should be used in addition to molecular tests. Hoffman
ity performance of the tests occurred for IgG in the group of et al evaluated an essay for COVID-19 IgG/IgM serological
symptomatic patients with more than 11 days after the onset rapid test cassette (17). The study included 29 PCR-confirmed
of symptoms. On the contrary, the worst performance occurred COVID-19 cases from a Swedish hospital and 124 negative
controls from a biobank. The authors reported a sensitivity of evolution. We calculated a pooled sensitivity of 76% (CI 71 – 80)
69% and 93.1% for IgM and IgG, and suggested that the test is and a pooled specificity of 98% (CI 97 – 98) for IgG. These results
not suitable for massive screening. Cassaniti et al reported the differ slightly from the reported by Castro et al in a metanaly-
results of a diagnostic essay carried out to evaluate the perfor- sis to estimate the pooled sensitivity and specificity in a group
mance of VivaDiag COVID‐19 IgM/IgG serological rapid test of 16 COVID-19 serological rapid tests available in Brazil (26).
in an emergency service in Italy (20). The study involved 110 The authors reported a sensitivity of 82% (CI 76 – 87) and a
subjects (30 positive for COVID-19). The authors reported a specificity of 97% (CI 96 – 98) for IgM, and a sensitivity of 97%
sensitivity of 18.4% and a specificity of 91.7%. A German study (CI 90 – 99) and specificity of 98% (CI 97 – 99) for IgG. Never-
aimed to evaluate the diagnostic capability of a rapid point-of- theless, the authors highlighted the high rate of false negative
care test for SARS-CoV-2 in a community screening (21). The results of these tests in the acute phase of the infection, which
study included 49 individuals (11 positive for SARS-CoV-2 con- may range between 10 – 44%. Also, they urged for a validation
firmed by RT-PCR). The authors reported a sensitivity of 36.4% study using Brazilian samples (26).
and a specificity of 88.9%). Ong et al compared the diagnostic In this sense, our study using samples from Colombian
accuracy of serological rapid tests in patients with suspected serums found that serological rapid tests should be limited to
coronavirus disease 2019 (22). The study evaluated five com- patients with more than 11 days of symptoms. Besides, the sero-
mercial rapid tests (Boson Biotech Rapid 2019-nCoV IgG/IgM logical tests are not an adequate strategy for the identification of
Combo Test Card; Cellex qSARS-CoV-2 IgG/IgM Cassette asymptomatic and pre-symptomatic patients. The recommen-
Rapid Test; Dynamiker Biotechnology 2019-nCOV IgG/IgM dations from the World Health Organization should be taken
Rapid Test; Orient Gene Biotech COVID-19 IgG/IgM Rapid Test into account; they should be avoided as a diagnostic test and
Cassette; and Prometheus Bio 2019-nCOV IgG/IgM Rapid) (22). they must be accompanied by RT-PCR (8).
It involved 228 samples from positive inpatients admitted in a This study has some limitations. First, in the absence of a gold
hospital in the Netherlands. Authors reported heterogeneous standard test, RT-PCR was used. Nucleic acid detection by the
sensitivities ranging from 10% (95% CI 0% – 23%) to 55% (95% Berlin protocol, although standardized and validated by the US
CI 33% – 77%). The test with the best performance was the Ori- National Institutes of Health, may have some false negatives,
ent Gene Biotech COVID-19 IgG/IgM Rapid Test (OGBRT) (22). especially during the first days of the clinical picture. On the
Consistent with the reported literature, our findings showed other hand, the pooled sensitivity analyses for IgM and IgG in
a lower performance for sensitivity in all the study groups. the asymptomatic and symptomatic groups showed moderate
These results differ from the reported sensitivity and specific- heterogeneities, which could invalidate the pooled sensitivity
ity values from commercial manufactures, which may occur data.
for many different reasons. First, most of the experiments in All in all, serological tests are essential to analyze the immune
the clinical reports from the manufactures did not report the response against infectious diseases to carry out their epidemi-
origin (country) used to evaluate the operative c apability of the ological characterization, natural resistance or susceptibility,
assays. The results may change from one country to other, or and the study of potential vaccines. In cases such as the SARS-
may be affected by conditions such as ethnicity and the time of CoV-2 pandemic, the secondary validation of the available
the onset of symptoms. Second, some of the information leaf- tests is essential to establish their usefulness as well as their
lets reported several samples used to estimate the sensitivity limitations. Our study showed that the usefulness of immuno-
and specificity, but all of them fail to report the clinical stage of chromatographic serological tests for COVID-19 is limited and
the patients. Besides, none of the tests reported the inclusion its use as a diagnostic tool has led to a large proportion of false
of serum from asymptomatic patients. This may result in an negatives promoting the contagion. Immunochromatographic
overestimation of the sensitivity of the test. Third, it should be tests may be used in seroprevalence studies once more sensi-
considered that most of the clinical reports from commercial ble antigens are incorporated in the technique, making them
manufacturers are not published in peer-reviewed journals, suitable tests for screening the infectious population.
were independent readers may evaluate the experiment con-
ditions. Lastly, it should be taken into account the diversity of Funding. This study was supported by the Instituto Nacional
the SARS-CoV-2 strains circulating in Colombia. Recent reports de Salud within the framework of the resources provided for
from the phylogenetic analysis of SARS-CoV-2 in Colombia the COVID-19 pandemic.
have shown that the lineages B.1.1, B.1.5 of SARS-CoV-2 are
the most common in the country (23,24). These lineages have Conflict of interests. None declared.
been reported in studies from Europe, the United States and
Australia (25). The described lineages may be linked with low Authors´ contributions. All authors have participated in the
performance of rapid diagnostic tests, as reported in Italian, design, analysis of results, writing and revision of the docu-
German, and Dutch studies (20–22). ment, and final approval of the manuscript.
We should mention the case of the Leccurate test, which per-
formance highly differs from the other 8 tests. Even when we Acknowledgements. The authors thank Professor Owen Har-
carried out a subgroup analysis considering the manufacturing rison for the revision of the manuscript; to Dr. Sofía Duque who
place (Asia vs Europe; Asia vs United States; China vs Europe; kindly reviewed the last version of the manuscript and pro-
China vs United States) the results of the pooled analysis vided valuable suggestions and comments.
increased the inconsistency. We did not find a reason to explain
the heterogeneity shown by the above mentioned test. Disclaimer. Authors hold sole responsibility for the views
Also, we have found that the best performance of the test expressed in the manuscript, which may not necessarily reflect
occurred in the group of patients with 11 or more days of the opinion or policy of the RPSP/PAJPH and/or PAHO.
REFERENCES
1. Zhou P, Yang X-L, Wang X-G, Hu B, Zhang L, Zhang W, et al. A Confidence Intervals in Large Meta-Analyses. Emmert-Streib F, edi-
pneumonia outbreak associated with a new coronavirus of probable tor. PLoS ONE [Internet]. 2012 Jul 25 [cited 2020 Jun 27];7(7):e39471.
bat origin. Nature. 2020 Mar;579(7798):270–3. Available from: https://dx.plos.org/10.1371/journal.pone.0039471
2. Udugama B, Kadhiresan P, Kozlowski HN, Malekjahani A, Osborne M, 16. Zamora J, Abraira V, Muriel A, Khan K, Coomarasamy A. Meta-
Li VYC, et al. Diagnosing COVID-19: The Disease and Tools for Detec- DiSc: a software for meta-analysis of test accuracy data. BMC
tion. ACS Nano [Internet]. 2020 Mar 30 [cited 2020 Apr 17]. Available Med Res Methodol [Internet]. 2006 Dec [cited 2020 Jun 16];6(1):31.
from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7144809/ Available from: https://bmcmedresmethodol.biomedcentral.com/
3. Musselwhite C, Avineri E, Susilo Y. Editorial JTH 16 –The Corona- articles/10.1186/1471-2288-6-31
virus Disease COVID-19 and implications for transport and health. 17. Hoffman T, Nissen K, Krambrich J, Rönnberg B, Akaberi D, Esmaeil-
J Transp Health [Internet]. 2020 Mar [cited 2020 May 6];16:100853. zadeh M, et al. Evaluation of a COVID-19 IgM and IgG rapid test; an
Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PM efficient tool for assessment of past exposure to SARS-CoV-2. Infect
C7174824/ Ecol Epidemiol [Internet]. 2020 Apr 14 [cited 2020 May 12];10(1).
4. di Giacomo E, Bellelli G, Peschi G, Scarpetta S, Colmegna F, Girol- Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/
amo G, et al. Management of older people during the COVID‐19 PMC7178815/
outbreak: Recommendations from an Italian experience. Int J Geri- 18. Delgado G, Vargas J, Mercado M, Gaviria P, Álvarez C. Toward to
atr Psychiatry [Internet]. 2020 Jul [cited 2020 Sep 24];35(7):803–5. establish selection criteria for rapid serological tests for COVID-19.
Available from: https://onlinelibrary.wiley.com/doi/abs/10.1002/ Infectio [Internet]. 2020 May 12 [cited 2020 Sep 24];24(3):17. Avail-
gps.5318 able from: http://www.revistainfectio.org/index.php/infectio/arti
5. Bachelet VC. Do we know the diagnostic properties of the tests cle/view/869
used in COVID-19? A rapid review of recently published literature. 19. Zhang J, Liu J, Li N, Liu Y, Ye R, Qin X, et al. Serological detection
Medwave [Internet]. 2020 Apr 28 [cited 2020 May 6];20(3). Available of 2019-nCoV respond to the epidemic: A useful complement to
from: /link.cgi/English/Reviews/GeneralReviews/7891.act nucleic acid testing. medRxiv. 2020 Mar 10;2020.03.04.20030916.
6. Long Q-X, Liu B-Z, Deng H-J, Wu G-C, Deng K, Chen Y-K, et al. 20. Cassaniti I, Novazzi F, Giardina F, Salinaro F, Sachs M, Perlini S,
Antibody responses to SARS-CoV-2 in patients with COVID-19. Nat et al. Performance of VivaDiag COVID‐19 IgM/IgG Rapid Test is
Med [Internet]. 2020 Apr 29 [cited 2020 May 6];1–4. Available from: inadequate for diagnosis of COVID‐19 in acute patients referring
https://www.nature.com/articles/s41591-020-0897-1 to emergency room department. J Med Virol [Internet]. 2020 Apr 8
7. Xiang F, Wang X, He X, Peng Z, Yang B, Zhang J, et al. Antibody [cited 2020 Sep 17];jmv.25800. Available from: https://onlinelibrary.
Detection and Dynamic Characteristics in Patients with COVID-19. wiley.com/doi/abs/10.1002/jmv.25800
Clin Infect Dis [Internet]. 2020 Apr 19 [cited 2020 May 6]:ciaa461. 21. Döhla M, Boesecke C, Schulte B, Diegmann C, Sib E, Richter E, et al.
Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/ Rapid point-of-care testing for SARS-CoV-2 in a community screen-
PMC7188146/ ing setting shows low sensitivity. Public Health [Internet]. 2020 May
8. World Health Organization. Advice on the use of point-of-care [cited 2020 Sep 17];182:170–2. Available from: https://linkinghub.
immunodiagnostic tests for COVID-19: scientific brief, 8 April 2020 elsevier.com/retrieve/pii/S0033350620301141
[Internet]. Geneva: World Health Organization; 2020. Available 22. Ong DSY, de Man SJ, Lindeboom FA, Koeleman JGM. Compar-
from: https://apps.who.int/iris/handle/10665/331713 ison of diagnostic accuracies of rapid serological tests and ELISA
9. Hajian-Tilaki K. Sample size estimation in diagnostic test studies to molecular diagnostics in patients with suspected coronavirus
of biomedical informatics. J Biomed Inform [Internet]. 2014 Apr 1 disease 2019 presenting to the hospital. Clin Microbiol Infect [Inter-
[cited 2020 May 7];48:193–204. Available from: http://www.sci- net]. 2020 Aug [cited 2020 Sep 17];26(8):1094.e7-1094.e10. Available
encedirect.com/science/article/pii/S1532046414000501 from: https://linkinghub.elsevier.com/retrieve/pii/S1198743X203
10. Vidal-Anzardo M, Solis G, Solari L, Minaya G, Ayala-Quintanilla B, 03050
Astete-Cornejo J, et al. Evaluación en condiciones de campo de una 23. Malagón-Rojas J, Mercado M, Ospina M, Parra E. Portadores asin-
prueba serológica rápida para detección de anticuerpos IgM e IgG tomáticos de SARS-CoV-2: cohorte de viajeros y trabajadores del
contra SARS-CoV-2. Rev Peru Med Exp Salud Pública [Internet]. aeropuerto internacional El Dorado de Bogotá‚ 2020 [Internet].
2020 Apr 21 [cited 2020 May 12];37(2). Available from: https:// XII Encuentro Nacional de Enfermedades infecciosas; 2020 Sep 25;
rpmesp.ins.gob.pe/index.php/rpmesp/article/view/5534 Cali, Colombia. Available from: https://www.acin.org/index.php/
11. Corman VM, Landt O, Kaiser M, Molenkamp R, Meijer A, Chu DK, acin/eventos
et al. Detection of 2019 novel coronavirus (2019-nCoV) by real-time 24. Laiton-Donato K, Villabona Arenas CJ, Usme Ciro JA, Franco Munoz
RT-PCR. Eurosurveillance [Internet]. 2020 Jan 23 [cited 2020 Jun C, Alvarez-Diaz DA, Villabona-Arenas LS, et al. Genomic epidemi-
27];25(3). Available from: https://www.ncbi.nlm.nih.gov/pmc/ ology of SARS-CoV-2 in Colombia [Internet]. Infectious Diseases
articles/PMC6988269/ (except HIV/AIDS); 2020 Jun [cited 2020 Sep 24]. Available from:
12. Silva Fuente-Alba C, Molina Villagra M. Likelihood ratio (razón de http://medrxiv.org/lookup/doi/10.1101/2020.06.26.20135715
verosimilitud): definición y aplicación en Radiología. Rev Argent 25. Rambaut A, Holmes EC, O’Toole Á, Hill V, McCrone JT, Ruis C,
Radiol [Internet]. 2017 Jul 1 [cited 2020 May 7];81(3):204–8. Avail- et al. A dynamic nomenclature proposal for SARS-CoV-2 lineages
able from: http://www.elsevier.es/es-revista-revista-argentina-rad to assist genomic epidemiology. Nat Microbiol [Internet]. 2020 Jul
iologia-383-articulo-likelihood-ratio-razon-verosimilitud-definicio 15 [cited 2020 Sep 24]. Available from: http://www.nature.com/
n-S0048761916301910 articles/s41564-020-0770-5
13. Aznar-Oroval E, Mancheño-Alvaro A, García-Lozano T, 26. Castro R, Luz PM, Wakimoto MD, Veloso VG, Grinsztejn B, Per-
Sánchez-Yepes M. Razón de verosimilitud y nomograma de Fagan: azzo H. COVID-19: a meta-analysis of diagnostic test accuracy of
2 instrumentos básicos para un uso racional de las pruebas del lab- commercial assays registered in Brazil. Braz J Infect Dis [Internet].
oratorio clínico. Rev Calid Asist [Internet]. 2013 Nov 1 [cited 2020 2020 Mar [cited 2020 Jun 27];24(2):180–7. Available from: https://
May 7];28(6):390–1. Available from: http://www.elsevier.es/es-re- linkinghub.elsevier.com/retrieve/pii/S1413867020300295
vista-revista-calidad-asistencial-256-articulo-razon-verosimilitud-n
omograma-fagan-2-S1134282X13000523
14. Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring incon-
sistency in meta-analyses. BMJ [Internet]. 2003 Sep 4 [cited 2020
Jun 16];327(7414):557–60. Available from: https://www.bmj.com/
content/327/7414/557
15. Thorlund K, Imberger G, Johnston BC, Walsh M, Awad T, Thabane Manuscript received on 23 July 2020. Revised version accepted for publication on
L, et al. Evolution of Heterogeneity (I2) Estimates and Their 95% 2 October 2020.
Palabras clave Betacoronavirus; infecciones por coronavirus; pruebas serológicas; diagnóstico; Colombia.