Abstract
Introduction
As of 30 April 2020, the novel betacoronavirus SARS-CoV-2 had infected more than 3 172 000 individuals, killing over 224 000 people and spreading to more than 200 countries. Italy was the most affected country in Europe and the third most affected in the world in terms of the number of cases. Therefore, the aims of this study are: (1) to estimate the prevalence of asymptomatic SARS-CoV-2-positive individuals among the general population of Verona; (2) to assess the accuracy (sensitivity, specificity and predictive values) of an ELISA serological test for the screening of SARS-CoV-2.
Methods and analysis
The study will be carried out on a random sample of subjects aged at least 10 years from the general population of Verona. Participants will undergo the measurement of vital parameters (oxygen saturation measured by oximeter, respiratory rate and body temperature detected by laser thermometer), the administration of a COVID-19-related symptoms questionnaire, the collection of a blood sample and a nasopharyngeal swab. Our evaluation will include the statistical technique of Latent Class Analysis, which will be the basis for the estimation of prevalence.
Ethics and dissemination
The study protocol has been approved by the Ethics Committee of Verona and Rovigo provinces on 15 April 2020 (internal protocol number 2641CESC). The study results will be submitted for publication in international, peer-reviewed journals and the complete dataset will be deposited in a public repository. Most relevant data will be made available to policy-makers as well as disseminated to stakeholders and to the community.
Keywords: public health, diagnostic microbiology, epidemiology, molecular diagnostics
Strengths and limitations of this study.
Study based on the random sample of a general population.
Very low estimation standard error (max 1.5%).
Simultaneous comparison of symptoms, real-time PCR test and ELISA serological test.
Results will depend on the response rate.
Introduction
As of 30 April 2020, the novel betacoronavirus SARS-CoV-2 had infected more than 3 172 000 individuals, killing over 224 000 people and spreading to more than 200 countries.1 Italy was at the time the most affected country in Europe and the third most affected in the world in terms of the number of cases. The epidemic was posing an extremely difficult challenge to healthcare establishments, health workers and to the general population. The identification of asymptomatic SARS-CoV-2-positive individuals is crucial in reducing the spread of the virus throughout the world. The frequency of such cases is unknown, despite asymptomatic cases regularly being referred to in the literature.2–4
Data from the cruise ship ‘Diamond Princess’ have shown that the percentage of asymptomatic SARS-CoV-2-positive cases among all passengers and crew members tested prior to disembarkation was about 50%. This rate has since been revised as 17.9.5
A similar study focusing on a Japanese population returning from China found the percentage of asymptomatic cases to be 33.3%.6
Finally, a small group of residents in a skilled nursing facility has also been screened: 13 of 23 residents were defined as asymptomatic, but 10 of these developed symptoms over the subsequent 7 days.7
Currently, SARS-CoV-2 detection is based on accepted and standardised molecular methods that require about 4 hours to carry out in our laboratory, although more rapid molecular tests are being made available. Samples are accumulating in many laboratories that are at risk of being overwhelmed. This causes further critical delays in managing SARS-CoV-2-positive cases and obtaining a definitive COVID-19 diagnosis.
Serological tests would be extremely useful for identifying asymptomatic carriers of SARS-CoV-2. Despite this, to date, the accuracy of these tests is insufficient to replace the current laboratory diagnosis. Serological tests focus on the detection of immunoglobulin (Ig)M and/or IgA and IgG. As it was noted during the previous SARS epidemic, a possible problem with serologic tests may be a cross-reaction with other coronaviruses.8 A recent publication has evaluated the median seroconversion time for antibodies IgA, IgM and IgG. The authors evaluated 173 SARS-CoV-2 -positive subjects and reported a median time of 11, 12 and 14 days, respectively. Additionally, antibodies were found in >40% of patients within 1 week of symptom onset, rapidly increasing to 100.0% (IgA), 94.3% (IgM) and 79.8% (IgG) from day 15 after symptom onset. In contrast, RNA detectability decreased from 66.7% (58/87) in samples collected before day 7 to 45.5% (25/55) during days 15–39. Combining RNA and antibody detection significantly improved the sensitivity of COVID-19 diagnosis (p<0.001), even in the early phase during the first week of symptoms (p=0.007). Moreover, a higher titre of the antibody was independently associated with a worse clinical classification (p=0.006).9 Therefore, the aims of this study are as follow:
To estimate the prevalence of asymptomatic SARS-CoV-2-positive individuals among the general population of Verona.
To assess the accuracy (sensitivity, specificity and predictive values) of two, commercially available serological tests for the screening of SARS-CoV-2.
Methods and analysis
Study design
This will be an observational cross-sectional prevalence study and an observational prospective diagnostic study in Verona (Veneto, Italy). Samples will be collected and analysed at Istituto di Ricovero e Cura a Carattere Scientifico(IRCCS) Sacro Cuore Don Calabria Hospital.
Study population
The study will be carried out on a random sample of the general population of Verona. Subjects eligible for the study must reside in Verona, be at least 10 years of age and provide consent for the participation in the study and for the donation of biological samples for study purposes.
Subjects will be randomly selected from the municipality of Verona’s registry and invited to participate.
According to official sources, the cumulative number of SARS-CoV-2 infections in Verona as of 25 May 2020 was 1528 cases (0.7% of the total population), of which 144 deaths, for a death rate of 9.4%.10
Procedures
Each randomly selected citizen will receive an invitation letter to their place of residence, outlining the reasons for the study and how to take part.
Those who decide to participate will be invited to contact a dedicated telephone number, at which point general information will be given and their contact details collected.
The specially trained staff at IRCCS Sacro Cuore Don Calabria Hospital will contact citizens both to confirm participation by verbal consent and to arrange an appointment (according to a pre-established calendar, automatically managed by a suitable software that limits the formation of queues). During the same phone call, a COVID-19-related symptoms questionnaire (online supplemental annex 1) will also be administered and all information related to the logistics and implementation of the study will be provided (eg, mask and gloves use, methods of sample collection and so on). This is to minimise the duration of physical contact and length of stay by participants in the centre, as well as to maximise protection against possible contagion.
bmjopen-2020-040036supp001.pdf (61.5KB, pdf)
In the case of participants under the age of 18, the phone call and questionnaire will be conducted with a parent (preferably the primary caregiver), whereas the minor will be required for the sample collection only.
Each participant that attends the IRCCS Sacro Cuore Don Calabria Hospital will be asked to deliver signed informed consent forms (customised according to the age of the participant), which will be verified and countersigned by the principal investigator or by delegated staff. Participants who fail to bring their consent form can obtain another copy directly at the centre.
Participants will then undergo the measurement of vital parameters (oxygen saturation measured by oximeter, respiratory rate and body temperature detected by laser thermometer), the collection of a blood sample and a nasopharyngeal swab. All procedures will be performed by specialised personnel of the IRCCS Sacro Cuore Don Calabria Hospital in dedicated clinical settings with suitable anticontagion equipment.
Auxiliary staff will monitor the movement of people outside and inside the hospital.
So as not to potentially contaminate hospital rooms and to avoid the need for continual sanitisation, samples will be taken outdoors in a designated tented facility located inside the hospital grounds, so as to allow examinations to be carried out even in the event of adverse weather. For those arriving by motorised vehicle, swabs may be collected from participants while they remain seated inside it, as a final step, when they are going out after concluding the other procedures including blood sampling.
Up to 280 samples per day are expected to be collected.
The collected samples will be stored in a refrigerator at 4°C until their transfer, within 24 hours, to the testing laboratories of the IRCCS Sacro Cuore Don Calabria Hospital. Blood samples will be immediately stored there on reception at −80°C, until their processing and analysis that will be carried out in the following weeks, whereas swabs are processed on reception and then also stored at −80°C.
Sample collection procedures will take place with 10 min intervals between the end of one sample collection and the start of the next one to ensure privacy and to avoid any close contact between subjects.
To further guarantee the safety of participants, the route inside the centre will be one-way only eliminating the need to return to spaces already frequented. Visual indications, such as strips on the ground, will allow all individuals to maintain the recommended safe distance from each other.
If a subject is unable to travel to the testing centre, then the home collection of the sample will be arranged during the initial phone call to confirm verbal consent.
The results of the examinations will be communicated to participants and in case of a positive SARS-CoV-2 result, appropriate procedures will be activated.
All essential information, including completed questionnaires and selected laboratory findings, will be recorded in an electronic Case Report Form using the platform OpenClinica.
Measurements
This protocol refers to STAndards for the Reporting of Diagnostic accuracy studies (STARD) guidelines11 for the reporting of diagnostic test accuracy. Based on an assessment methodology already used at the IRCCS in diagnostic studies,12 13 the assessment will be carried out using an approved molecular test as the gold standard, assuming that the sensitivity will not be 100% (due to variances in nasopharyngeal swabbing technique). The evaluation will also include the statistical technique of latent class analysis (LCA), which will be also the basis for the estimation of prevalence.
Anti-SARS-CoV-2 ELISA IgA/IgG (Euroimmun, Germany) will be performed according to the manufacturer’s instructions, detecting SARS-CoV-2 antibodies of classes IgA (described as early markers of acute respiratory tract infections) and IgG (indicating a persisting or past infection). In a recent study,14 the value of specific IgA detection, in the early detection of acute SARS-CoV-2 infections, has been confirmed. The assay uses the S1 domain of the spike protein on the surface of SARs-CoV-2 as its antigen, which is considered to be more specific for the serological detection of SARS-CoV-2 antibodies.
The SARS-CoV-2 IgG assay (Abbott Laboratories, USA) is a chemiluminescent microparticle immunoassay for the detection of IgG antibodies to SARS-CoV-2. This will be also performed according to the manufacturer’s instructions.
According to the most recently published study on both tests,15 the sensitivity and specificity of the Euroimmun test were found to be 78.3% and 96.7% for IgG, and 86.7% and 82.7% for IgA; for the Abbott test (IgG), they were 81.8% and 99.3%, respectively. Another study in preprint16 found (for IgG) a sensitivity of Euroimmun test varying from 76.9% to 87.1%, and for Abbott from 96.2% to 97.1%, depending on the population group sampled; specificity was 97% for Euroimmun and 99% for Abbott. These figures were obtained on small groups of patients, some of whom with active or recent infection when they might have not yet developed detectable antibodies.
The Primary Reference Standard test is a real-time reverse transcription PCR (RT-PCR), executed at our department, that has been set up according to the procedures followed by the Regional Reference Laboratory (Department of Microbiology, University Hospital of Padua) and cross-validated.
RNA will be extracted from nasopharyngeal swabs in accordance with our routine laboratory practice. Briefly, total RNA will be extracted using a MagnaPure LC.2 instrument (Roche Diagnostic, Monza, Italy), and MagNA Pure LC RNA Isolation Kit-High Performance (Roche), according to the manufacturer’s instructions for the cell-containing samples. Eluted RNA will be analysed following the routine in-house real-time RT-PCR protocol for the COVID-19 diagnostic test. The remaining RNA aliquots will be stored at −80°C until they are required for further tests.
True positive subjects will be those with a positive real-time RT-PCR result as this indicates the presence of SARS-CoV-2 RNA. True negative subjects will be subjects with a negative real-time RT-PCR result that have no detectable SARS-CoV-2 RNA. The outcome for indeterminate results is outlined later.
In detecting SARS-CoV-2, RT-PCR revealing specific viral RNA sequences may be considered the gold standard, being a test with virtually 100% specificity and therefore acceptable as a gold standard for the sensitivity of index tests. However, the sensitivity of RT-PCR cannot truly be considered to be 100% in detecting SARS-CoV-2 RNA, as is becoming clear to reference laboratories performing the test, due to the viral load being too low for the sequences to be revealed or a flawed swabbing technique. In cases that only use this gold standard, the classification of discordant results (negative gold standard, positive index test) would be subject to error. Using a composite reference standard (CRS) is one of the alternative methods when a ‘perfect’ gold standard is not available.17 18 However, this method has its limitations too, as when a CRS is used, its accuracy cannot be assumed ‘a priori’.19 Alternative methods to address a lack of a gold standard are LCMs.20 LCA is planned using the available tests for SARS-CoV-2 as well as other, selected, clinical and paraclinical variables.
Each test will be executed independently by experienced laboratory personnel. Laboratory professionals will not be aware of the clinical data of the subjects and will not know in advance the results of any other test.
Subjects found positive
Subjects with a positive test result will be informed of the test result and managed according to the routine procedures for clinical assessment and isolation.
Sample size calculation
To the best of our knowledge, there is no information published in an accredited scientific journal that indicates the prevalence of asymptomatic SARS-CoV-2-positive individuals among the general population. In Italy, epidemiologists have reported a potential prevalence equal to 5 or 10 times higher than the number of detected SARS-CoV-2-positive individuals. Other similar sources indicate that the prevalence of asymptomatic SARS-CoV-2-positive subjects is 9%–10% of the general population.21 22
Therefore, assuming an objective population of 235 034 inhabitants (residents in the municipality of Verona on 1 January 2020 who are aged at least 10 years—source Istituto nazionale di statistica - ISTAT), a prevalence of asymptomatic SARS-CoV-2-positive individuals of 10.0% and an alpha value (type 1 error) of 5%, then a random sample of 1527 subjects is required to obtain a SE of no more than 1.5%, that is, a 95% CI of lower amplitude or equal to 3%. Assuming a drop-out rate of 35%, it will be necessary to enrol 2061 subjects.
A systematic probabilistic sampling technique will be used to perform the sample list.
Data analysis plan
Demographic and clinical data will be summarised using descriptive statistics, measures of variability and precision, and plots. Statistical tests will be used based on the type of variables, test assumptions and sample dimension. All parameters will be reported with 95% CIs. Statistical models and estimations will be adjusted for covariates if necessary. Information criteria and likelihood ratio methods will be used to compare candidate models.
Test results will be displayed in contingency tables of which sensitivity, specificity and predictive values will be calculated. Estimations will be reported along with the 95% CIs.
For the LCA, we will use latent class models (LCMs) to boost the diagnostic accuracy of the gold standard test by adding to these models the results of multiple tests used in the same samples, as well as other key information from the CRF.
The basis of LCA is that each subject belongs to one of a finite number of classes; in our study, we have two classes: with and without COVID-19. Each class is described by a set of parameters that define the statistical distribution of outcomes; here the conditional probability is that the subject has or does not have the condition COVID-19 (specificity and sensitivity) and the probability that the condition COVID-19 is present (prevalence).20 Observations with missing reference standard results will be excluded from the analysis.
Discussion and conclusions
Currently, there are no published data that reliably estimate the prevalence of asymptomatic SARS-CoV-2-positive individuals in Italy. SARS-CoV-2 cases are now reported worldwide but at differing incidences depending on the region. Developed regions with a temperate climate and a medium to high population density seem to be the most affected. However, the true prevalence of asymptomatic SARS-CoV-2-positive individuals is unknown, as is the prevalence of those who have never contracted the virus. Furthermore, it is not yet established whether now recovered, previously SARS-CoV-2-positive individuals can become reinfected.
This study will be useful in understanding the spread of SARS-CoV-2 in a city with a total population of around 260 000 inhabitants and a population density of approximately 1300 inhabitants per square kilometre (ISTAT source). Most importantly, the prevalence of asymptomatic SARS-CoV-2-positive individuals and the prevalence of those who are negative for SARS-CoV-2 can be estimated. This will allow phase II of Italy’s outbreak management strategy, in which day-to-day activities will gradually be reintroduced and population contact will resume, to be planned more effectively.
In addition, the study will allow us to better understand the symptoms of SARS-CoV-2 infection and what potential role they may have in disease prognosis. Finally, the study will allow us to evaluate the diagnostic accuracy and value of the ELISA serological test as a screening tool for the general population.
Ethics and dissemination
This protocol will be registered at http://www.clinicaltrials.gov. The full study protocol will be made accessible at a public repository on publication.
This study adheres to the National Health and Medical Research Council National Statement on Ethical Conduct in Human Research and the principles of the Declaration of Helsinki.
The study could involve vulnerable groups within the community, and so it is imperative that the study is conducted in a sensitive and culturally appropriate manner. On an invitation to the study, subjects will be given the opportunity to review all study materials and ask any questions. Furthermore, subjects who feel overwhelmed or anxious at any point during study participation will be referred to an appropriate support service. Subjects will also be reassured that they are free to withdraw from the study at any time without reason or consequence. Results from the study will be disseminated through a presentation at national and international conferences and publications in peer-reviewed journals.
Patient and public involvement statement
Patient and public involvement (PPI) representatives worked with us to refine the research question, however it was difficult to involve patients in other areas of the study design due to the very technical methods required to do a data linkage analysis. PPI representatives will write a plain language summary and design a leaflet for dissemination to their peers and distributing them to patient groups.
Ethics approval
This study was reviewed and approved by the CESC (Comitato Etico per le Sperimentazioni Cliniche—approval number 28, 17 April 2020).
Supplementary Material
Acknowledgments
The authors thank Nicoletta De Santis for the editing, Elinor Julie Rae Anderson for language editing and Elvia Malo that contributed to the preparation of documents for ethical approval.
Footnotes
Contributors: MG, CP conceived the study. MG, CP, ZB, AP and CM contributed to study design. MG, CP and ZB contributed to the preparation of the manuscript; preparation of documents for ethical approval including study materials (questionnaire, participant information statement, etc). All authors have contributed to the review of manuscript and are directly involved in the study scientific committee.
Funding: This work was supported by the Italian Ministry of Health “Fondi Ricerca Corrente - Linea 3” to IRCCS Sacro Cuore Don Calabria Hospital.
Competing interests: None declared.
Patient consent for publication: Not required.
Provenance and peer review: Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
References
- 1.World Health Organization’s COVID-19 situation report 64- 24/03/2020. Available: https://www.who.int/docs/default-source/coronaviruse/situation-reports/20200324-sitrep-64-covid-19.pdf?sfvrsn=703b2c40_2 [Accessed 25 Mar 2020].
- 2.Chan JF-W, Yuan S, Kok K-H, et al. . A familial cluster of pneumonia associated with the 2019 novel coronavirus indicating person-to-person transmission: a study of a family cluster. Lancet 2020;395:514–23. 10.1016/S0140-6736(20)30154-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Mizumoto K, Kagaya K, Zarebski A, et al. . Estimating the asymptomatic proportion of coronavirus disease 2019 (COVID-19) cases on board the diamond Princess cruise ship, Yokohama, Japan, 2020. Euro Surveill 2020;25:2000180 10.2807/1560-7917.ES.2020.25.10.2000180 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Wei M, Yuan J, Liu Y, et al. . Novel coronavirus infection in hospitalized infants under 1 year of age in China. JAMA 2020;323:1313. 10.1001/jama.2020.2131 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.National Institute of Infectious Diseases Field Briefing: diamond Princess COVID-19 cases, 2020. Available: https://www.niid.go.jp/niid/en/2019-ncov-e/9417-covid-dp-fe-02.html
- 6.Nishiura H, Kobayashi T, Miyama T, et al. . Estimation of the asymptomatic ratio of novel coronavirus infections (COVID-19). medRxiv 2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kimball A, Hatfield KM, Arons M, et al. . Asymptomatic and Presymptomatic SARS-CoV-2 Infections in Residents of a Long-Term Care Skilled Nursing Facility - King County, Washington, March 2020. MMWR Morb Mortal Wkly Rep 2020;69:377–81. 10.15585/mmwr.mm6913e1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Lee HK, Lee BH, Seok SH, et al. . Production of specific antibodies against SARS-coronavirus nucleocapsid protein without cross reactivity with human coronaviruses 229E and OC43. J Vet Sci 2010;11:165–7. 10.4142/jvs.2010.11.2.165 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Zhao J, Yuan Q, Wang H, et al. . Antibody responses to SARS-CoV-2 in patients of novel coronavirus disease 2019. Clin Infect Dis 2020. doi: 10.1093/cid/ciaa344. [Epub ahead of print: 28 Mar 2020]. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Azienda zero – Regione del Veneto, SARS-CoV-2 in Veneto. Available: https://www.azero.veneto.it/-/emergenza-coronavirus [Accessed May 2020].
- 11.Cohen JF, Korevaar DA, Altman DG, et al. . STARD 2015 guidelines for reporting diagnostic accuracy studies: explanation and elaboration. BMJ Open 2016;6:e012799. 10.1136/bmjopen-2016-012799 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Buonfrate D, Sequi M, Mejia R, et al. . Accuracy of five serologic tests for the follow up of Strongyloides stercoralis infection. PLoS Negl Trop Dis 2015;9:e0003491. 10.1371/journal.pntd.0003491 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Beltrame A, Guerriero M, Angheben A, et al. . Accuracy of parasitological and immunological tests for the screening of human schistosomiasis in immigrants and refugees from African countries: an approach with latent class analysis. PLoS Negl Trop Dis 2017;11:e0005593. 10.1371/journal.pntd.0005593 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Okba NMA, Müller MA, Li W, et al. . Severe acute respiratory syndrome coronavirus 2-specific antibody responses in coronavirus disease patients. Emerg Infect Dis 2020;26:1478–88. 10.3201/eid2607.200841 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Nicol T, Lefeuvre C, Serri O, et al. . Assessment of SARS-CoV-2 serological tests for the diagnosis of COVID-19 through the evaluation of three immunoassays: two automated immunoassays (Euroimmun and Abbott) and one rapid lateral flow immunoassay (NG biotech). J Clin Virol 2020;129:104511. 10.1016/j.jcv.2020.104511 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Strömer A, Grobe O, Rose R, et al. . Diagnostic accuracy of six commercial SARS-CoV-2 IgG/total antibody assays and identification of SARS-CoV-2 neutralizing antibodies in convalescent sera. medRxiv 2020. [Google Scholar]
- 17.Rutjes AWS, Reitsma JB, Coomarasamy A, et al. . Evaluation of diagnostic tests when there is no gold standard. A review of methods. Health Technol Assess 2007;11:iii, ix-51. 10.3310/hta11500 [DOI] [PubMed] [Google Scholar]
- 18.Reitsma JB, Rutjes AWS, Khan KS, et al. . A review of solutions for diagnostic accuracy studies with an imperfect or missing reference standard. J Clin Epidemiol 2009;62:797–806. 10.1016/j.jclinepi.2009.02.005 [DOI] [PubMed] [Google Scholar]
- 19.Dendukuri N, Schiller I, de Groot J, et al. . Concerns about composite reference standards in diagnostic research. BMJ 2018;360:j5779. 10.1136/bmj.j5779 [DOI] [PubMed] [Google Scholar]
- 20.Goetghebeur E, Liinev J, Boelaert M, et al. . Diagnostic test analyses in search of their gold standard: latent class analyses with random effects. Stat Methods Med Res 2000;9:231–48. 10.1177/096228020000900304 [DOI] [PubMed] [Google Scholar]
- 21.La Vecchia C. Doxa: potrebbero essere oltre 5 milioni gli italiani infettati da COVID-19 a marzo 2020. La Statale News Universit degli Studi di Milano, 2020. http://lastatalenews.unimi.it/doxa-potrebbero-essere-oltre-5-milioni-italiani-infettati-covid-19-marzo-2020 [Google Scholar]
- 22.Natale GD. The COVID-19 infection in Italy: a statistical study of an abnormally severe 2 disease. medRxiv 2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
bmjopen-2020-040036supp001.pdf (61.5KB, pdf)