SARS-CoV-2 Persistent Viral Shedding in the Context of Hydroxychloroquine-Azithromycin Treatment

SARS-CoV-2 nasopharyngeal shedding contributes to the spread of the COVID-19 epidemic. Among 3271 COVID-19 patients treated at the Hospital University Institute Méditerranée Infection, Marseille, France from 3 March to 27 April 2020, tested at least twice by qRT-PCR, the median SARS-CoV-2 nasopharyngeal shedding duration was 6 days (range 2–54 days). Compared with short shedders (qRT-PCR positivity < 10 days), 34 (1.04%) persistent shedders (qRT-PCR positivity ≥ 17 days; mean ± SD: 23.3 ± 3.8 days) were significantly older, with associated comorbidities, exhibiting lymphopenia, eosinopenia, increased D-dimer and increased troponin (p < 0.05), and were hospitalized in intensive care unit in 17.7% vs. 1.1% of cases (p < 0.0001). Viral culture was positive in six persistent shedders after day 10, including in one patient after day 17, and no viral co-pathogen was detected in 33 tested patients. Persistent shedders received azithromycin plus hydroxychloroquine ≥ 3 days in 26/34 (76.5%) patients, a figure significantly lower than in short shedders (86.6%) (p = 0.042). Accordingly, mortality was 14.7% vs. 0.5% (p < 0.0001). Persistent shedding was significantly associated with persistent dyspnea and anosmia/ageusia (p < 0.05). In the context of COVID-19 treatment, including treatment with azithromycin plus hydroxychloroquine, the persistence of SARS-CoV-2 nasopharyngeal shedding was a rare event, most frequently encountered in elderly patients with comorbidities and lacking azithromycin plus hydroxychloroquine treatment.


Introduction
SARS-CoV-2, responsible for COVID-19, is detected in the nasopharynx, which could constitute a major portal of entry for this emerging pathogen [1]. Indeed, not only is SARS-CoV-2 RNA routinely detected in nasopharyngeal swabs of COVID-19 patients [2], but this clinical material yields living viruses after it is appropriately cultivated in cell culture systems [3,4]. These data suggest that SARS-CoV-2 nasopharyngeal shedding is of interest for the natural history of COVID-19 in patients and in populations, as it may relate to the prognosis of the infection and its contagiousness. Two meta-analyses, including 79 and 28 studies, converged to indicate a viral shedding duration of 17 days (mean) and 18.4 days (median), respectively [5,6].
Monitoring more than 3200 COVID-19 patients firmly documented by real-time RT-PCR (qRT-PCR) at the Hospital University Institute Méditerranée Infection, Marseille, France in March-April 2020 provided the opportunity to clarify the clinical and virological characteristics of patients with persistent nasopharyngeal shedding of SARS-CoV-2 [2].

Patients
This retrospective study aimed to describe the duration of SARS-CoV-2 viral shedding among 3737 qRT-PCR-confirmed COVID-19 patients, followed from 3 March to 27 April 2020 in the Méditerranée Infection Institute, as previously reported [2]. Clinical, radiological and laboratory data were collected, as previously reported [2,7], and the severity of patient illness was evaluated using the National Early Warning Score version 2 (NEWS-2) and the Charlson score [8]. All patients without contraindications were offered oral hydroxychloroquine (HCQ) (200 mg, three times a day for ten days) and azithromycin (AZ) (500 mg at day 1, followed by 250 mg per day for 4 days), and ceftriaxone or ertapenem was added for seven days in case of a NEWS-2 ≥ 5 [9,10]. Initiation of medical care (i.e., medical consultation and initiation of treatment) was defined as day 0, outpatient reevaluation was offered at days 2, 6 and 10 (more if needed), while hospitalized patients were evaluated twice daily. For hospitalized patients, qRT-PCR testing was performed on respiratory samples collected every day after diagnosis. For ambulatory followed-up patients, respiratory samples were collected at days 2, 6, 10 after diagnosis, then every four days when still qRT-PCR-positive at day 10. Biological results used for the analysis were those of tests performed on the day of admission to hospital (day-care hospital or complete hospitalization). The main criteria for discharge from complete hospitalization were the mean (or a significant decrease in flow) of oxygen supplementation

Virology
All patients were SARS-CoV-2-diagnosed based on at least one positive qRT-PCR test performed by nasopharyngeal swab. The time of follow-up was defined relative to the qRT-PCR diagnosis. SARS-CoV-2 genome sequencing was performed directly from nasopharyngeal swab RNA extract using the Illumina MiSeq sequencer and Nextera XT paired-end technology (Illumina, Evry, France) and genome sequences were compared with the reference SARS-CoV-2 isolate Wuhan Hu-1, genome sequence (NC 045512.2) [11]. In case of qRT-PCR cycle threshold (Ct) value >18, partial spike gene PCR amplification (nucleotides 21,296-23,424 in reference to NC 045512.2) was performed, as previously described (DOI: https://doi.org/10.35088/4y1e-ec62, accessed on 11 May 2021). Viral cultures were performed as previously described [3,4]. Briefly, 500 µL of nasopharyngeal swab fluid (Virocult, Elitech, France) or sputum sample were passed through a 0.22-µm pore sized centrifugal filter (Merck millipore, Darmstadt, Germany), then inoculated in 4 wells of 96-well culture microplates containing Vero E6 cells (ATCC CRL-1586) into Minimum Essential Medium culture medium with 4% fetal calf serum and 1% glutamine. All samples were kept for less than 24 h at +4 • C before processing. After centrifugation at 4000× g, microplates were incubated at 37 • C. They were observed daily for evidence of cytopathogenic effect. Two subcultures were performed weekly. Presumptive detection of virus in supernatant showing cytopathic effect was done using the SU5000 scanning electron microscope (Hitachi High-Tech Corporation, Tokyo, Japan), before confirmation by a qRT-PCR targeting the SARS-CoV-2 gene E [2]. Culturing was performed systematically for hospitalized patients, and prospectively or retrospectively for non-hospitalized persistent viral shedders. Patients for whom the qRT-PCRs became negative within 10 days were designated as short viral shedders, patients with positive qRT-PCRs between day 10 and 17 as long viral shedders and patients with positive qRT-PCRs after day 17 as persistent viral shedders.

Co-Pathogen Detection
Co-pathogens in the nasopharyngeal swabs were screened by using a multiplex test incorporating 21 targeted pathogens, following the instructions of the supplier (FTD Respiratory pathogens 21; Siemens Fast Track Diagnosis, Luxembourg).

Statistical Analysis
We used the Student t-test, Mann-Whitney U test, Chi2 test or Fisher's exact test to compare differences between short viral shedding and persistent viral shedding groups. To explore risk factors associated with the duration of viral shedding, we performed multivariable analyses using linear regression models. All variables significant at p < 0.10 in univariate analyses were introduced in the initial multivariable model. A backward approach was then used to assess the iteration of variables and to control potential confounders (the significance level value for stay was set at 0.05). A primary model was performed over the full sample and a secondary model was performed among patients aged 65 years and older [12]. A secondary analysis using a multivariable logistic regression model was also performed to identify the risk factors associated with prolonged COVID-19 symptoms. The same backward approach was applied to this secondary outcome. A two-sided α of less than 0.05 was considered statistically significant. All analyses were carried out using SAS 9.4 statistical software (SAS Institute, Cary, NC, USA).

Ethics Statement
Data were collected retrospectively from the routine care setting. This non-interventional retrospective study was approved on 13 May 2020 by our institutional review board committee (Méditerranée Infection No.: 2020-021). In compliance with European General Data Protection Regulation No. 2016/679, patients were informed of the potential use of their medical data and that they could refuse the use of their data. The analysis of collected data followed the reference methodology MR-004 registered on No. MR 5010010520 in the AP-HM register, in compliance with European General Data Protection.

Care and Treatment
A higher percentage of persistent shedders was hospitalized in the intensive care unit compared to short shedders, 17.7% vs. 1.1% (p < 0.0001). In this cohort, 26/34 (76.5%) persistent shedders had HCQ-AZ therapy ≥ 3 days, versus 2426/2800 (86.6%) short shedders (p = 0.042, Mantel-Haenszel Chi-2 test). Indeed, 8/34 persistent shedders had HCQ-AZ therapy < 3 days: more precisely, one patient refused HCQ and AZ treatment, and seven other patients received AZ only because HCQ was contraindicated due to cardiac contraindications (n = 6) and drug interaction (n = 1) ( Table 1). In contrast, 374/2800 short shedders had HCQ-AZ therapy < 3 days (Table 3). Persistent viral shedding was also associated with persistent dyspnea 30 days after the onset of the follow-up. Mortality was 14.7% among persistent shedders vs. 0.5% among short shedders (p < 0.0001). At thẽ 9-10-month follow-up in January 2021, 5/34 persistent shedders had died and seven had been seen in consultation: three patients had persistent post-COVID-19 dyspnea, including one patient who had persistent ageusia. Kinetics analysis showed that eosinophil and lymphocyte counts were significantly lower in patients with persistent viral shedding than in those with short viral shedding, whereas neutrophils were higher in persistent viral shedders. D-dimers and troponin were significantly and persistently higher in persistent viral shedders (Supplementary Table S1).

Viral Genotype Analysis
A total of 21 complete (n = 11) or partial (approximately the first half of the spikeencoding gene, n = 10) genome sequences were obtained for the 34 persistent shedders, comprising 14 classified in Nextstrain clade 20A, 3 in clade 20B and 4 in clade 20C [13]. This distribution into viral clades did not differ from that observed in non-persistent viral shedders during the same period (310 in clade 20A, 55 in clade 20B and 94 in clade 20C, for a total of 466 genome sequences). All spike sequences harbored the D614G substitution that was present in almost all SARS-CoV-2 genomes in Europe since the epidemic onset. No additional mutations were present within the spike.

Viral Culture
A total of 177 nasopharyngeal swabs were cultured in the 34 patients, of which 69 swabs collected in 20 patients were positive in culture. More precisely, eight swabs collected in six patients had a positive culture when collected >10 days after the beginning of follow-up, at days 11, 13, 14, 15, 17 and 33 ( Figure 1). since the epidemic onset. No additional mutations were present within the spike.

Viral Culture
A total of 177 nasopharyngeal swabs were cultured in the 34 patients, of which 69 swabs collected in 20 patients were positive in culture. More precisely, eight swabs collected in six patients had a positive culture when collected > 10 days after the beginning of follow-up, at days 11, 13, 14, 15, 17 and 33 (

Discussion
In a large series of qRT-PCR-documented COVID-19 cases followed at the Méditerranée Infection Institute in March and April 2020, the median SARS-CoV-2 nasopharyngeal shedding was six days, shorter than that reported in the literature [14]. Moreover, less than 1% of cases (34/3737) exhibited shedding ≥17 days, and persistent positivity of qRT-PCR in nasopharyngeal swabs correlated with culture positivity. Testing procedures and discharge criteria remained the same during the study period. We observed positive cultures more than 10 days after the onset of the follow-up in eight of 34 patients, a situation rarely reported, as nasopharyngeal shedding is usually reported on the sole basis of qRT-PCR positivity [6]: one case of 4-month viral shedding was reported in the absence

Discussion
In a large series of qRT-PCR-documented COVID-19 cases followed at the Méditerranée Infection Institute in March and April 2020, the median SARS-CoV-2 nasopharyngeal shedding was six days, shorter than that reported in the literature [14]. Moreover, less than 1% of cases (34/3737) exhibited shedding ≥17 days, and persistent positivity of qRT-PCR in nasopharyngeal swabs correlated with culture positivity. Testing procedures and discharge criteria remained the same during the study period. We observed positive cultures more than 10 days after the onset of the follow-up in eight of 34 patients, a situation rarely reported, as nasopharyngeal shedding is usually reported on the sole basis of qRT-PCR positivity [6]: one case of 4-month viral shedding was reported in the absence of culture and contagiousness [15], whereas anecdotal persistent viral shedding, with positive culture for up to 60 days, has been reported in cancer patients receiving chemotherapy [16].
Previous case reports and a small series of patients experienced prolonged shedding with SARS-CoV-2 (Table 4) [17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35], mostly involving immunocompromised individuals, including patients diagnosed with hematological malignancies in most cases.  This was not our situation, where only one patient had lymphoma and another had immunosuppression following kidney transplantation. In addition, several cases of viral shedding beyond 90 days occurred in patients who received treatments with convalescent plasma and/or remdesivir [17,18]. These situations were not encountered in this series, as no patient received remdesivir or convalescent plasma and the figures reported here were observed in the context of standardized care, including the prescription of hydroxychloroquine and azithromycin treatment. It is noteworthy that almost one quarter of the 34 persistent shedders reported here did not receive the combination of hydroxychloroquine and azithromycin, a prevalence significantly higher than that in short shedders. This standardized care notably differed in other series reporting a higher proportion of persistent shedders, yet populations may not have been similar, bringing into question the impact of combination of hydroxychloroquine plus azithromycin at the dosage prescribed in reducing the time of SARS-CoV-2 nasopharyngeal shedding.
This reported series mirrors the first COVID-19 epidemic in our region, which was caused by SARS-CoV-2 of clades 20A, 20B and 20C. Accordingly, based on next-generation population sequencing we did not observe any genotype pattern in any of the 21 patients whose viral genome could be explored. Further ultradeep sequencing of paired early and late nasopharyngeal swabs in the 34 patients reported here may nevertheless reveal quasi-species and minority genotypes that may have escaped the standard genotyping methods used in this report.
Whether the data here reported on the genetic context of SARS-CoV-2 circulating in our region one year ago would apply to that of SARS-CoV-2 variants responsible for current epidemics [36] remains to be explored.

Informed Consent Statement:
In compliance with European General Data Protection Regulation No. 2016/679, patients were informed of the potential use of their medical data and that they could refuse the use of their data. The analysis of collected data followed the reference methodology MR-004 registered on No. MR 5010010520 in the AP-HM register, in compliance with European General Data Protection. Data Availability Statement: Data are available from the corresponding author upon reasonable request.

Conflicts of Interest:
The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.