In December 2019, during the early stages of the coronavirus disease 2019 (COVID-19) outbreak, discharge guidelines for hospitalized patients were not clear, leading to many patients remaining hospitalized until their real-time reverse transcription-polymerase chain reaction (rRT-PCR; hereinafter referred to as PCR) tests returned negative results. The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) PCR test is crucial for diagnosing COVID-19 [1, 2], monitoring the disease progression [3, 4], and determining whether hospitalized patients can be discharged [5, 6]. However, interpreting the PCR result can be more challenging during follow-up than at the time of diagnosis. During the initial diagnosis, a large amount of virus particles is present in the test specimen, making it easy to determine positive or negative results. However, during follow-up testing, especially at the time of discharge, the virus concentration is low, rendering interpretation of the results difficult.

Since the emergence of COVID-19 in 2019, numerous studies have reported on the diagnosis [7, 8], clinical features [9, 10], and treatment of the disease [11, 12]. Additionally, the implementation and clinical application of the PCR-based SARS-CoV-2 detection assay have been well reported [13-15]. However, few studies have conducted follow-up PCR testing on SARS-CoV-2-positive patients and analyzed the results. Therefore, the aim of our present study was to analyze changes in the follow-up PCR results obtained from the repeated testing of hospitalized COVID-19 patients.

The diagnostic and follow-up data used in this study were from COVID-19 patients in Daegu, the site of the first large-scale infection in Korea. During the study period, COVID-19 diagnosis was conducted using the SARS-CoV-2 rRT-PCR assay, making analysis of the results challenging. There were instances where it was difficult to determine the results at the time of monitoring and discharge. By analyzing the follow-up PCR results of SARS-CoV-2-positive patients in detail, we aim to provide useful information that can be used to improve the diagnosis and monitoring of COVID-19 or any new infectious diseases.

This study was approved by the Institutional Review Board of Daegu Catholic University Medical Center. We analyzed the PCR results of patients diagnosed with COVID-19 at Daegu Catholic University Medical Center from March to May 2020. In total, 84 patients who underwent four or more PCR tests and had their medical records confirmed were selected for the study. The PowerCheck 2019-nCoV Kit (Kogene Biotech, Seoul, Korea) and CFX96 Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA) were used for the SARS-CoV-2 PCR tests, and delta cycle threshold (ΔCt) values and clinical features were retrospectively analyzed.

A Ct value of 39 was used as the cut-off for determining positive and negative PCR results. PCR negativity was defined as the first days when negative results were received twice in a row [16]. If either the E or RdRp gene alone was positive, the results were considered positive when the positive results persisted and negative when the negative results persisted. For patients who tested negative, a subsequent positive and then negative result was determined on the basis of both the E and RdRp genes being positive and negative, respectively. To observe changes in the Ct values (i.e., ΔCt ≥8.0), the results for 20 patients with a PCR Ct value of 30 or less at the time of initial diagnosis were evaluated for changes based on the duration of hospitalization.

Statistical analyses were performed using MedCalc software (version 20.218; MedCalc Software Ltd., Ostend, Belgium). Oneway analysis of variance was used to analyze the differences in ΔCt values among the follow-up PCR subgroups. The Wilcoxon signed-rank test was used to compare the Ct values of the E and RdRp genes. The Mann–Whitney test was used to compare time intervals for negative conversion of the SARS-CoV-2 PCR results. A P-value of less than 0.05 was considered statistically significant.

1. Demographic data and SARS-CoV-2 PCR results

Demographic data and SARS-Cov2 rRT-PCR results are summarized in Table 1. Of the 84 patients (median age, 67 years), 50 were females and 34 were males. At the time of initial diagnosis, their median Ct values for the E and RdRp genes were 33.3 and 33.6, respectively. These patients underwent nine repeat PCR tests over a span of 38.5 days, obtaining a negative result after a median of 28.0 days. Among the 84 patients, 77 tested negative and 7 remained PCR positive throughout the study period (Fig. 1). Of the 77 patients with negative results, 58 remained negative (consistently negative group), while 19 later tested positive before subsequently returning to a negative status (re-positive group). The Ct values of the E and RdRp genes at the time of re-positivity were 34.7 and 35.2, respectively, which were higher than the values recorded at the initial diagnosis (Table 1).

Figure 1. Time interval of SARS-CoV-2 rRT-PCR positivity status of 84 COVID-19 patients. Data are presented as median days (d).

Table 1 Summary of follow-up SARS-CoV2 rRT-PCR results from repeated testing of hospitalized patients

Parameters	Results
Total number	84
Age	67.0 (52.0–77.5) years
M:F	34:50
Initial Ct value
E gene	33.3 (29.0–35.1)
RdRp gene	33.6 (29.5–35.9)
Duration for negative conversion	28.0 (20.0–37.0) days
No. of repeat rRT-PCRs	9.0 (5–13)
rRT-PCR follow-up duration	38.5 (26.5–49.5) days
No. of negative conversion	58 (69.0%)
No. of Neg>Pos>Neg (re-positive)	19 (22.6%)
No. of persistent positivity	7 (8.3%)
Ct value of re-positive group
E gene	34.7 (31.9–36.8)
RdRp gene	35.2 (32.4–37.0)

Data are presented as the number (%) or median (interquartile range, Q1–Q3).

Abbreviations: rRT-PCR, real-time reverse transcription-PCR; Neg, negative; Pos, positive.

2. Changes in Ct values based on the duration of hospitalization

Table 2 presents the median ΔCt values for the E and RdRp genes based on the duration of hospitalization. The ΔCt value was calculated by subtracting the present Ct value from the initial value obtained on the duration of hospitalization. The ΔCt values of the two genes showed a continuous increase from 5.16 to 12.6 (E gene) and from 5.15 to 12.02 (RdRp gene) until the 25th day of hospitalization, indicating a decline in the gene concentrations over time. Notably, however, the ΔCt values of both genes started to decrease after the 25th day.

Table 2 Median ΔCt values of the E and RdRp genes after the initial diagnosis

Duration, days (Group No.)	No. of patients	ΔCt E	ΔCt RdRP	95% CI ΔCt E	95% CI ΔCt RdRP	ΔCt E difference (P <0.05)	ΔCt RdRP difference (P <0.05)
2–5 (1)	9	5.16	5.15	2.727–7.602	2.625–7.669	(2)(3)(4)(5)(6)(7)(8)	(2)(3)(4)(5)(6)(7)(8)
6–10 (2)	16	7.97	7.92	5.961–9.969	6.001–9.848	(1)(3)(4)(5)(6)(7)(8)	(1)(3)(4)(5)(6)(7)(8)
11–15 (3)	19	10.57	10.55	8.272–12.875	8.120–12.988	(1)(2)(4)(5)(6)(7)(8)	(1)(2)(5)(8)
16–20 (4)	15	10.80	10.66	7.740–13.868	7.310–14.006	(1)(2)(3)(5)(6)(7)(8)	(1)(2)(5)(6)(7)(8)
21–25 (5)	17	12.60	12.02	9.744–15.456	9.207–14.842	(1)(2)(3)(4)(6)(7)(8)	(1)(2)(3)(4)(6)(7)(8)
26–30 (6)	10	11.69	10.96	8.127–15.247	7.521–14.405	(1)(2)(3)(4)(5)(7)(8)	(1)(2)(4)(5)(7)(8)
31–40 (7)	20	12.07	11.31	10.033–14.105	9.261–13.349	(1)(2)(3)(4)(5)(6)(8)	(1)(2)(4)(5)(6)(8)
41–50 (8)	12	9.56	8.91	6.707–12.410	6.089–11.726	(1)(2)(3)(4)(5)(6)(7)	(1)(2)(3)(4)(5)(6)(7)

Abbreviations: ΔCt, delta cycle threshold; E, envelope; RdRp, RNA-dependent RNA polymerase; CI, confidence interval.

3. Concordance rates of PCR results for the E and RdRp genes

The concordance rate of the PCR test results for the E and RdRp genes was 82.4%, with both genes being positive in 424 samples (60.2%) and negative in 156 samples (22.2%) (Table 3). However, the PCR test discrepancy rate between the two genes was 17.6%, with 8.1% of samples testing positive for the E gene and negative for the RdRP gene and 9.5% testing positive for the RdRp gene and negative for the E gene.

Table 3 Concordance rates of the rRT-PCR results for the E and RdRp genes

RdRp gene	E gene		Total
	Positive	Negative
Positive
Number (%)	424 (60.2)	67 (9.5)	491
Ct value (mean±SD)	E (33.6±4.1) RdRp (34.0±4.2)	38.2±0.1
Negative
Number (%)	57 (8.1)	156 (22.2)	213
Ct value (mean±SD)	37.6±1.1	Not available
Total	481	274	704

Abbreviations: rRT-PCR, real-time reverse transcription PCR; E, envelope; RdRp, RNA-dependent RNA polymerase.

4. Distribution of Ct values for the E and RdRp genes

Among the 424 positive samples, the median Ct values of the E and RdRp genes were 34.43 and 35.07, respectively, with the higher Ct value for the RdRP gene being statistically significant (P<0.001) (Fig. 2).

Figure 2. Comparison of the Ct values (median and interquartile range) of the E and RdRp genes for the 424 SARS-CoV-2-positive samples (Box and whisker plot, P<0.0001 by the Wilcoxon signed-rank test).

5. Comparison of the time intervals for negative PCR conversion between the negative and re-positive groups

The time interval for negative PCR conversion was compared between the consistently negative and re-positive groups. The mean number of negative conversion days for the negative and re-positive groups were 27.8 and 33.7 days, respectively, with the longer period for the re-positive group being statistically significant (P<0.05) (Fig. 3).

Figure 3. Comparison of the time intervals (days, mean±2 SD) for negative SARS-CoV2 rRT-PCR conversion between the consistently negative and re-positive groups (P=0.049 by the independent t-test).

6. Analysis of the patients with persistent SARS-CoV-2-positive PCR results

Of the 84 patients in the study, seven tested persistently positive for SARS-CoV-2. Among these seven patients, three succumbed to the illness, one required transfer to an intensive care center, and three showed improvement and were subsequently transferred to an isolated observation center (Table 4). In the four patients with a poor prognosis, the levels of C-reactive protein (CRP) remained elevated. Conversely, in the three patients who showed improvement, the CRP levels decreased significantly. Notably, among the four patients with initial Ct values of approximately 25 (Patients 2, 4, 6, and 7), the two with a poor prognosis exhibited only a slight change in Ct values (2.5–3.9 cycles for patients 4 and 7), whereas the other two with a favorable prognosis showed a substantial change in Ct values (10–12.2 cycles for patients 2 and 6).

Table 4 Summary of rRT-PCR data and clinical outcomes of patients with persistent SARS-CoV2-positive results

Patients	Age (years)	Sex	rRT-PCR (+) period (day)	Initial rRT-PCR Ct (E/RdRp)	No. of rRT-PCR repeats	Final rRT-PCR Ct (E/RdRp)	Summary of rRT-PCR repeats	Initial CRP (mg/dL)	Final CRP (mg/dL)	Clinical outcome
Patient 1	63	M	35	30.7/31.4	9	33.3/34.6	4/9 only RdRp (+)	82.0	103.0	Deceased (multiorgan failure)
Patient 5	77	M	41	35.1/34.4	10	38.5/38.5	7/10 both genes (–)	52.0	70.1	Deceased (ARDS)
Patient 7	76	F	45	25.6/25.8	6	28.6/28.3	6/6 both genes (+)	37.9	117.0	Deceased (ARDS)
Patient 4	58	M	52	25.8/26.6	12	29.7/29.4	1/12 both genes (–)	13.6	155.0	Transferred to KNUH for intensive therapy
Patient 2	86	M	49	24.1/24.8	18	34.4/37.0	2/18 both genes (–)	126.0	17.9	Transferred to DMC for maintenance
Patient 3	87	F	42	38.0/38.6	10	39.3/38.2	1/10 both genes (–)	14.7	2.0	Transferred to DMC for maintenance
Patient 6	82	M	61	24.4/25.0	20	35.1/35.0	1/20 both genes (–)	83.9	7.2	Transferred to DMC for maintenance

Abbreviations: rRT-PCR, real-time reverse transcription-PCR; CRP, C-reactive protein; ARDS, acute respiratory distress syndrome; KNUH, Kyungpook National University Hospital; DMC, Daegu Medical Center.

To address the lack of data on the ideal frequency and duration of PCR testing for COVID-19 patients, we analyzed the follow-up PCR records of 84 SARS-CoV-2-positive patients to evaluate the optimal duration and frequency of repeat testing. Challenges include discrepancies between the PCR results of the two genes tested and difficulties in adhering to the cut-off Ct values for positive results as suggested by manufacturers.

In this study, in which a total of nine tests were performed at intervals of 4.3 days over a period of 38.5 days, the median duration required for negative PCR conversion was 28.0 days. The median duration of SARS-CoV-2 detection in the upper respiratory tract has been reported to be 14.5–34.0 days [16-19]. Generally, the viral load of SARS-CoV-2 is very high on the day of symptom development and then gradually decreases, with the virus typically being undetectable within 3 weeks of symptom onset [16, 17, 20]. The duration of virus detection also depends on the chosen cutoff Ct value [19].

In the ΔCt analysis based on the duration of hospitalization, the viral concentration decreased rapidly until the second week of hospitalization and then decreased gradually thereafter, but it then surprisingly increased again after the fourth week. In clinically stable patients, the increase in viral concentrations after the fourth week does not indicate active viral replication but rather the presence of residual nucleic acid or persistent inactive viral particles. Therefore, repeated PCR testing for SARS-CoV-2 after 4 weeks may not accurately reflect the actual viral load in the patient.

Among the 77 patients who tested negative for SARS-CoV-2, 19 (22.6%) experienced PCR re-positivity (Fig. 1). Moreover, the median duration of the first virus-negative period was significantly longer in the re-positive group than in the consistently negative group (37.0 vs. 27.5 days). Wong et al. [21] reported that 34.3% of 35 patients showed SARS-CoV-2 re-positivity, without COVID-19 symptoms, likely as a result of residual viral nucleic acid. However, Hossain et al. [18] found that two of 12 patients with symptom recurrence also had a re-positive PCR result. Intermittent viral shedding has also been identified as a major cause of PCR re-positivity [22]. Hu et al. [23] reported that delayed clearance of SARSCoV-2 RNA was associated with an older age and chest tightness in hospitalized patients, whereas Abrahim et al. [24] found that patients with comorbidities, such as cancer and HIV infection, had a prolonged viral clearance period. Lim et al. [25] suggested that the duration of viral shedding was longer in individuals with more clinical symptoms, indicating that the time required for negative conversion might be related to the severity of clinical symptoms.

Compared with the 58 consistently virus-negative patients, the 19 re-positive patients were significantly older (median age 61.5 vs. 81.0 years), whereas other clinical symptoms were generally unremarkable. Among the 19 re-positive patients, 12 had CRP levels within the normal range, and two showed slightly increased levels (5.0–12.0 mg/dL). Seventeen of the re-positive patients recovered from COVID-19, whereas the remaining two patients experienced severe respiratory symptoms and died from respiratory failure. The re-positivity in the 17 recovered patients was likely due to residual nucleic acid or intermittent viral shedding. By contrast, the re-positivity in the two deceased patients may have been due to delayed viral clearance or persistent viral presence, indicating the potential ongoing complications of COVID-19.

The Ct values of the RdRp gene were higher than those of the E gene, aligning with the findings from another study that used the same PCR reagent [13]. This discrepancy in Ct values is attributed to variations in PCR conditions rather than differences in target gene concentration. Furthermore, the PCR concordance rate between the two genes was 82.4%, reflecting a mismatch rate of 17.6% (Table 3). Sung et al. [13] reported that in follow-up PCRs of confirmed infected patients, only one of the target genes might test positive and the detected gene may vary.

Of the seven patients who remained persistently virus positive, three experienced symptom improvement, whereas the other four individuals either died or required intensive care. In a study comparing 67 patients, Chang et al. [4] found that patients who tested persistently positive for SARS-CoV-2 had higher incidences of intensive care requirement and pulmonary infiltration. Conversely, Ikegami et al. [20] reported that six out of 272 recovered patients continued to test positive for SARS-CoV-2 in nasopharyngeal swab PCR assays. This indicates that persistent viral detection can occur in both recovered and deteriorating patients. Therefore, the clinical significance of PCR results has to be interpreted in conjunction with CRP levels or ΔCt values.

In summary, it took a median of 28.0 days for 77 patients to test negative in the SARS-CoV-2 PCR assay, with the re-positive group taking a median of 37.0 days to achieve this. An 82.4% concordance rate between the E and RdRp genes was observed. We conclude that for clinically stable patients, continued follow-up PCR testing for up to 4 weeks is advisable to assess negative PCR conversion.

None declared.