key: cord-0857974-m12poscd
authors: COVISAL Guyane,; nacher, m.; demar, m.
title: Prospective comparison of saliva and nasopharyngeal swab sampling for mass screening for COVID-19
date: 2020-09-24
journal: nan
DOI: 10.1101/2020.09.23.20150961
sha: d7f95999be643bb1e097e34f5cb902e9b9626456
doc_id: 857974
cord_uid: m12poscd

Current testing for COVID-19 relies on quantitative reverse-transcriptase polymerase chain reaction from a nasopharyngeal swab specimen. Saliva samples have advantages regarding ease and painlessness of collection, which does not require trained staff and may allow self-sampling. We enrolled 776 persons at various field-testing sites and collected nasopharyngeal and pooled saliva samples. 162 had a positive COVID-19 RT-PCR, 61% were mildly symptomatic and 39% asymptomatic. The sensitivity of RT-PCR on saliva samples versus nasopharygeal swabs varied depending on the patient groups considered or on Ct thresholds. There were 10 (6.2%) patients with a positive saliva sample and a negative nasopharyngeal swab, all of whom had Ct values<25. For symptomatic patients for whom the interval between symptoms onset and sampling was <10 days sensitivity was 77% but when excluding persons with isolated Ngen positivity (54/162), sensitivity was 90%. In asymptomatic patients, the sensitivity was only 24%. When we looked at patients with Cts <30, sensitivity was 83% or 88.9% when considering 2 genes. The relatively good performance for patients with low Cts suggests that Saliva testing could be a useful and acceptable tool to identify infectious persons in mass screening contexts, a strategically important task for contact tracing and isolation in the community.

Current testing for COVID-19 relies on quantitative reverse-transcriptase polymerase chain 39 reaction (RT-qPCR) from a nasopharyngeal swab specimen.(1) Nasopharyngeal sampling 40 requires human resources and training, personal protective equipment and swabs, and time, 41 generating testing bottlenecks and potential exposure to transmission at crowded testing 42 sites. Moreover, the unpleasantness of the procedure and the long waiting delays may 43

dissuade some persons to get tested or to repeat tests when they are negative. There is an 44 urgent need for innovative testing strategies to rapidly identify cases, reduce waiting delays, 45

and facilitate mass screening. Saliva samples have advantages regarding ease and 46 painlessness of collection, which does not require trained staff and may allow self-sampling. 47

The comparison of real time PCR results on salivary and nasopharyngeal samples has shown 48 discrepancies between studies, with most finding greater sensitivity and lower RT-PCR Cts 49

in nasopharyngeal swab samples(2-4) whereas others found greater sensitivity in saliva 50 samples(5, 6). The sources of variation may have been the study population (hospitalized 51 patients versus screening of contacts or mildly symptomatic patients), saliva collection 52 techniques and timing, conditioning and delays in processing raw saliva samples, or 53 differences in the RT-PCR techniques used. 54

We here report the first prospective study of the performance of saliva testing compared to 55 nasopharyngeal swabs in a field context of mass screening in French Guiana. 56

Context 58

This French territory neighboring Amapa state in Brazil has been highly impacted by 59 COVID-19 with 3.2% of the population having had a confirmed infection, notably among the 60 poorest populations.(7) In this context, testing and tracking were implemented throughout 61 the epidemic, testing tents and mobile testing teams including the remote health centers, the 62

Red Cross, Médecins du Monde, and the reinforcements from the Réserve Sanitaire were 63 coordinated by the regional health agency to investigate around clusters of cases. The testing 64

efforts for this small population peaked to nearly 0.5% of the population screened in a day. 65

Between July 27 th and September 10 th , we prospectively enrolled consecutive, persons aged 3 67 years or more with mild symptoms suggestive of COVID-19 and high-risk asymptomatic 68

persons at various testing sites and mobile testing brigades in French Guiana reaching 69 remote sites up to 240 km in the Amazonian Forest. During screening missions, mobile 70 teams, consisting of Healthcare personnel (doctors, nurses) were coordinated by the Health 71

Regional Agency of French Guiana, targeting villages, neighborhoods, where the virus was 72

circulating collected persons often out of doors or in health centers. These mobile teams were 73 made up of staff from the Red Cross, Médecins du Monde, the Cayenne hospital PASS and 74 the health reserve. Team travel was coordinated and decided by the health regional agency 75

of French Guiana each week during a weekly update and was guided the knowledge of 76 clusters or a screening campaign in different neighborhoods. Inclusion criteria were: males or 77

females with an indication to perform a COVID diagnostic test (symptomatology, contact 78 case, systematic screening, etc.), aged at least 3 years old. Non-inclusion criteria were refusal 79 of the patient or his legal representative, person taking treatments that reduce salivary 80 volume (anticholinergic activity), impossibility of carrying out the NPS, and persons under 81 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 24, 2020. . guardianship or curatorship, or placed under protective measures. All study participants 82

were enrolled and sampled in accordance with the protocol. An investigator explained the 83 objectives of the study and obtain the consent of the patient or his legal representative. The 84

form was completed by the investigator or a person delegated by the investigator. The 85

trained nurse present during the testing mission performed the nasopharyngeal swab 86

collected the salivary sputum sample in a urine container. A trained agent carried out a short 87

questionnaire. At the end of each day, all completed forms and samples were sent to 88

Cayenne hospital and stored at 4°C before analysis. Samples and participant information 89

were non-individually identifiable and collected with a unique identifying number. 90

The same technique was used for the 2 samples throughout the study: the QIAsymphony  92 and GeneFinder kit, a Real-time PCR assay. GeneFinder TM COVID-19 detects SARS-CoV-2 by 93 amplification of RdRp gene, E gene and N gene according to WHO's recommended protocol. 94

Viral nucleic acid was extracted by using the QIAamp DSP viral kit on the QIAsymphony 95 RGQ, an integrated fully automated nucleic acid extraction (chemical lysis and paramagnetic 96

bead binding) and sample preparation platform (Qiagen GmbH, Germany). The real-time 97

PCR assays for SARS Cov2 were performed with an Applied 7500 cycler (Thermofisher) with 98

the Genefinder kit (Ellitech group) that could detect the Ngen, RdRP and E gen. As the 99

Nucleic acid extraction methods could affect the results of viral nucleic acid amplification 100 tests, we treated the couple saliva-nasopharyngeal specimens with the same method and 101 most of the time in the same series, the eluates were obtained from 200µl of specimens (300 102 µL minus 100 µL dead volume). The remainder of each sample was divided into paired 103 aliquots kept in a biorepository for further studies evaluating new screening tools. 104

Statistical analysis was performed using STATA® 16 (Stata corporation, College Station, 106

Texas, USA). Cross tabulations considering different subgroups was performed. We 107 considered the specific genes for SARS-Cov2 RdRp and Ngene to calculate different Ct 108

categories. 109

The protocol received ethical approval from the Comité de Protection des Personnes under 111 the number 2020-A02009-30/SI:20.07.07.54744. 112

We included 776 patients, 162 (20.9%) of whom had a positive result (152 nasopharyngeal 114 and 86 saliva) (Figure 1 ). The sex ratio (M/F) was 1.6, the mean age was 40 (standard 115 deviation=16.8). Overall, 61% were mildly symptomatic and 39% were asymptomatic. 116

For symptomatic patients, 84% had a symptoms onset <10 days, and 4% were hospitalized 117 within 2 weeks after inclusion. The sensitivity of RT-PCR on saliva samples versus 118 nasopharygeal swabs varied depending on the patient groups considered (Fig 2a) or on Ct 119 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 24, 2020. . thresholds (Fig 2b) . There were 10 (6.2%) patients with a positive saliva sample and a 120

negative nasopharyngeal swab, all of whom had Ct values<25. For symptomatic patients for 121

whom the interval between symptoms onset and sampling was <10 days sensitivity was 77% 122 but when excluding persons with isolated Ngen positivity (54/162), sensitivity was 90%. In 123 asymptomatic patients, the sensitivity was only 24% (Fig 2a) . Recent studies have argued 124

that transmission potential -estimated by the capacity to infect cell cultures-was restrained 125

to those with low Cts (8, 9) , a proxy for high viral load. When we looked at patients with Cts 126 <30, sensitivity was 83% or 88.9% when considering 2 genes. Figure 3 shows a trend for 127

fanning towards the higher Ct values the nasopharyngeal versus saliva sample scatterplots 128

for the different genes amplified by RT PCR. The positive predictive value of saliva samples 129 was 88.4% and the negative predictive value was 88.9%. 130

Contrarily to 2 studies suggesting a greater positivity rate for saliva(5, 6), we observed that 132 saliva testing was less sensitive than nasopharyngeal swabs. Whereas most studies were 133

hospital-based collecting saliva in the early morning before mouth rinsing and breakfast, our 134 study was a screening study that was performed in difficult field conditions targeting hard 135

to reach populations after breakfast and teeth brushing, moreover out of doors in a tropical 136

context. The poor sensitivity on asymptomatic positive nasopharyngeal swabs was 137

presumably also linked to the inclusion of non-infectious patients in the denominator. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 24, 2020. . samples. There was a "fanning" pattern with greater dispersion at higher Ct values for 160 different genes in the nasopharyngeal and saliva samples. 161

Laboratory Diagnosis of COVID-19: 164 Current Issues and Challenges

Evaluating the use of posterior oropharyngeal saliva in a point-of-care assay for the 167 detection of SARS-CoV-2

Saliva sampling for diagnosing SARS-CoV-2 infections in symptomatic patients and 170 asymptomatic carriers

Saliva as a non-invasive 172 specimen for detection of SARS-CoV-2

Comparing nasopharyngeal swab and early morning saliva for the identification of 175 SARS-CoV-2

Saliva or Nasopharyngeal Swab Specimens for 178 Detection of SARS-CoV-2