Prolonged SARS‐CoV‐2 viral shedding in patients with chronic kidney disease – O’Sullivan – – Nephrology

Currently, clinical diagnosis of SARS‐COV‐2 is made primarily using a viral throat swab and subsequent positive polymerase chain reaction (PCR) result over a pre‐specified CT threshold. Initial interim World Health Organization (WHO) guidance recommended confirming clearance of SARS‐CoV‐2 in clinically recovered patients by demonstrating two negative quantitative reverse transcription PCR (RT‐PCR) results on sequential nasopharyngeal swabs taken >24 hours apart.¹ While the link between PCR swab positivity and infectious risk via viral shedding has been questioned,² recent WHO guidance has provided pragmatic guidance acknowledging a potential role for sequential negative swabs in high‐risk populations or where the infective risk to others is high.³

Immune dysfunction is well‐recognized in chronic kidney disease (CKD). Furthermore, patients with CKD may have previously received immunosuppressive therapies, or have additional risk factors such as hypertension that could be associated with a delay in viral clearance. Certainly, persistent of SARS‐CoV‐2 has been noted in immunosuppressed hosts.⁴ A recent review of 79 studies of SARS‐CoV‐2 found a mean SARS‐CoV‐2 RNA shedding duration in upper respiratory tract of 17 days but none described patients with CKD.⁵ We hypothesized that this altered immune milieu in CKD may lead to delayed clearance of the SARS‐CoV‐2 virus, prolonged RT‐PCR positivity and longer inpatient stay for patients awaiting negative nasopharyngeal swabs. This would have important clinical implications for service provision, discharge planning and communicating expectations and care pathways with our patients.

1 METHODS

This cross‐sectional observational study involved patients admitted with SARS‐CoV‐2 infection at the Queen Elizabeth University Hospital Glasgow and Edinburgh Royal Infirmary, Scotland, UK between March 1, 2020 and May 31, 2020. Exemption from the formal ethical review was granted as data were routinely collected during normal clinical care and were aggregated and anonymized. Local data‐protection policies and best practice were followed per Caldicott guidance.⁶

Patients were identified through routinely collected health‐care data. Patients over the age of 18 years were eligible for inclusion if admitted from March 1, 2020 to May 31, 2020 and had a positive viral swab for SARS‐CoV‐2 within 48 hours of admission. All patients who remained in hospital until two consecutive negative swabs were obtained prior to discharge were included. Swabs were tested at most every 48 hours once free from symptoms as per local protocol.

These are typically patients in nursing homes, residential care homes, or deemed to be an infective risk on discharge to vulnerable individuals.

Patients requiring admission to intensive care unit or high dependency unit were excluded (leaving only level 0 and level 1 patients whose entire admission was spent on medical wards). Patients who did not survive to discharge were excluded. Patients with acute kidney injury (AKI), or who required new renal replacement therapy were excluded as were renal transplant recipients and kidney donors.

Patients with CKD were identified and stratified according to kidney disease: improving global outcomes (KDIGO) CKD stage based on modification of diet in renal disease estimated glomerular filtration rate (eGFR). Each patient’s acute and historical renal function was reviewed by a nephrologist. Patient’s electronic health records were consulted in ambiguous cases to understand the context of any historical eGFR measurements. The first eGFR considered was the admission eGFR during the SARS‐CoV‐2 infection. If the eGFR was >60 mL/min/1.73 m² and remained so over the admission (ie, no AKI) this patient was considered not to have CKD. Where an eGFR was 60 mL/min/1.73 m² or less, we reviewed all of their available eGFR on that admission to confirm whether the value was rapidly changing and representing AKI. Next, historical eGFR values were considered. When more than one value was available these were considered in the broader clinical context of the patient. For example, if they were consistent with each other over time, this was felt to be stable CKD. If progressive decline was noted and the admission eGFR was consistent with the trajectory, we assigned CKD according to the most recent value.

For the categorical classification of CKD, stages 1 and 2 (eGFR > 60) were considered ‘no CKD’ as per KDIGO.

SARS‐CoV‐2 infection severity scores were recorded as mild, moderate, severe or critical as per definitions proposed by Qiu et al.^{7, 8} Comorbidities were recorded including cardiovascular disease, hypertension, diabetes, cancer and HIV status.

The primary outcome was time in days from admission to reach the first of two consecutive negative nasopharyngeal swabs for SARS‐CoV‐2 infection. SARS‐CoV‐2 positivity was established using a number of platforms in routine clinical use using RT‐PCR assays to detect the RdRP gene and N gene, the E gene or the ORF1a/b and E gene. Multiple platforms allowed for internal re‐testing and cross‐validation in the case of uncertain results.

Difference between groups was tested using the Mann‐Whitney test, and a P‐value ≤.05 was considered statistically significant. Comparisons between CKD stages were performed with Kruskal‐Wallis Rank Sum Test and pairwise Wilcox test using the Holm correction for multiple hypothesis testing. Comparisons of proportions (eg, comparing sex between CKD status) was performed using a Pearson’s chi‐squared test. Multivariable linear regression was performed using the finalfit package.⁹ Multivariable linear regression analysis was conducted using the explanatory variables of age, sex, disease severity, comorbidities including cardiovascular disease, hypertension, diabetes, cancer and either the presence of CKD, CKD stage or eGFR. All statistical analyses were performed using R version 4.

2 RESULTS

A total of 138 patients were admitted with SARS‐CoV‐2 infection and met inclusion criteria. Of them, 80 (58%) patients had CKD. The mean age overall was 73 years (SD = 14). Patients with CKD were older with a mean age of 76.3 (SD = 11.9) as compared to patients without, who had a mean age of 69.1 (SD = 16, P =  .008). Patient characteristics are shown in Table 1. Overall, 71 (51%) patients were female, although there were a higher proportion of female patients in the CKD group (37% female in a group without CKD vs 61% female in a group with CKD, P = .014). Patients with CKD were more likely to have hypertension, cardiovascular disease and have mild SARS‐CoV‐2 infection. Notably, no patients had HIV infection. The mean weekly swabs were 2.31 (range: 0.2‐4.2) in the control group and 2.15 (range: 0.5‐4.2) in the CKD group (P = .58).

The median time from admission to the first of two consecutive negative nasopharyngeal swabs was 18 days (range = 1‐44) in patients with CKD, compared with 11 days (range: 1‐71) in patients without CKD (P = .001) (Figure 1A).

CKD, CKD stage or eGFR were independently associated with prolonged time to viral clearance, whereas all other variables were not (Figure 1B, Table 2). For every decrement in eGFR of 10 mL/min/1.73 m², time to clearance increased by approximately 1 day, accounting for 2.5% of the variation in time to clearance in the data.

3 DISCUSSION

Our data suggest that patients with CKD who are admitted to hospital with SARS‐CoV‐2 take longer to achieve sequential negative nasopharyngeal swab RT‐PCR results than those without CKD.

Early data suggested viral shedding lasted until approximately 20 to 31 days after symptom onset.^10–12 Unfortunately, there were few patients with CKD in these cohorts. However, in 10 renal transplant recipients in Wuhan, viral shedding was indeed longer than a control group (28.4 ± 9.3 vs 12.2 ± 4.6 days).¹² Of course, these patients are exogenously immunosuppressed and unlikely to be representative of the CKD cohort more broadly. Factors previously associated with prolonged viral shedding in cohorts without CKD were male sex, increased age, hypertension, delayed admission, severe illness, mechanical ventilation and corticosteroid treatment.¹³ Interestingly, despite higher rates of hypertension and cardiovascular disease in patients with CKD, these were not associated with longer time to clearance in our dataset, suggesting that CKD may supersede them as risk factors for delayed clearance when included in analysis, or that these factors are less important when critically ill mechanically ventilated patients are excluded. An understanding length of duration of viral shedding, as it relates to disease severity, is lacking. A recent meta‐analysis found 13 of 20 studies which had recorded disease severity found increased shedding with more severe disease, but the remain studies found either no change or reduced duration.⁵ Itis important to note that the presence of RT‐PCR positivity on nasopharyngeal swab does not imply infectivity but does have a significant impact on clinical pathways and length of stay when negative swabs are required prior to discharge. Sequential negative tests nonetheless remain a tool in clinical decision making and risk assessment.

There are important limitations to note. There is potential for lead time bias if patients with CKD present earlier to hospital, although this would have resulted in an apparent decrease in time to clearance. We have no record of the duration of acute illness—a longer duration of symptomatic illness in patients with CKD could delay viral clearance or testing. Our control group was likely to have a similar disease severity as all patients were sick enough to require admission, but remained on medical wards and survived to discharge, as well as sharing the requirement to remain in hospital awaiting negative tests suggesting similar social situations and a minimum shared level of frailty. As discussed in the methods, the sickest of patients were excluded to avoid severe disease confounding results. However, the data did not allow for matching for disease severity, multimorbidity or associated risk factors that may have contributed to CKD which may have independently contributed to delayed viral clearance, such as hypertension. This is an important when considering the mechanism of our findings. While the pragmatic interpretation of results would remain unchanged—that is, CKD is associated with longer delays to negative swabs, confounders of this nature would alter the underlying pathophysiological explanation. It may not be CKD per se that results in the additional delay. Nonetheless, the label of CKD captures these risk factors on the wards and we hope will prompt more in‐depth research in future to identify specific mechanisms. The prevalence of CKD is high at 9.1% of the global population.¹⁴ As such, any delay to viral clearance in such a major cohort, regardless of underlying aetiology, has important implications for patients and clinicians caring for them.

In conclusion, our findings suggest a prolonged time to viral clearance in patients with CKD, with implications for renal service provision, discharge planning and hospital capacity as well as a direct impact on patients due to extended hospital stay, potential anxiety and stigmatisation.

CONFLICTS OF INTEREST

The authors have no conflicts of interest to declare.