Different interpretations of inconclusive results of SARS-CoV-2 real-time RT PCR

Abstract

We investigated whether inconclusive results could be interpreted differently depending on the situation. First, data from retesting of the initial samples from subjects without a confirmed COVID-19 history were analyzed. And by analyzing the results of consecutive tests with new specimens after receiving inconclusive results between arrivals and locals for 2 periods. As a result, 179 of 219 cases (81.7%) showed still inconclusive or weakly positive results. If contamination is well controlled in a general laboratory, the effectiveness of retesting with the same sample is limited. The rate of the subsequently positive patient was significantly higher in locals than in arrivals and periods with a higher positive rate. The inconclusive results could be interpreted differently depending on the epidemiologic background and the positive rate at that time.

1. Introduction

Since the first case of coronavirus disease (COVID-19) was reported in December 2019, variants ranging from Alpha, Beta, Gamma, Delta, and Omicron have occurred, resulting in more than 52 million confirmed patients as of May 2022 [1,2]. Real-time reverse transcription polymerase chain reaction (rRT PCR) testing has been introduced to detect the SARS-CoV-2 virus, which causes COVID-19, and is being used for diagnosis [3].

In initial guidelines, the WHO recommended that at least 2 SARS-CoV-2 targets are detected as a positive result, and at least 1 of 2 genes should be specific to the SARS-CoV-2 virus [4]. The guidelines recommended by the Korea Disease Control and Prevention Agency (KDCA) and the Korean Society for Laboratory Medicine (KSLM) in Korea also require the use of molecular tests targeting 2 or more regions of the SARS-CoV-2 gene. In addition, when all targets are amplified, the test is determined to be positive [5,6].

The inconclusive result of amplifying only a part of the target genes in the SARS-CoV-2 rRT PCR test is commonly observed in patients recovering from COVID-19 infection. Still, mutations in the primers or probe sites of the target gene, instability of the reagent itself, or cross-contamination of the sample or amplification product can also cause an inconclusive result [6], [7], [8]. For this reason, when an inconclusive result is obtained in a laboratory, retesting is performed with the same sample according to the manufacturer's guidelines. If the retest result is the same, it is classified as an inconclusive result, and a new sample is recommended to be collected and tested because an inconclusive result in a patient without a confirmed COVID-19 history might mean an incubation period or an early stage of infection [6]. Inconclusive results not only delay the diagnosis of COVID-19 but also retesting with the same sample or performing consecutive tests with a new sample are significant concerns of the laboratory since overall resources required for the testing process are reintroduced.

As a private referral laboratory, we have experienced many inconclusive results. In this study, data from retesting of the initial samples from subjects without a confirmed COVID-19 history were first analyzed. By analyzing the results of consecutive tests with new specimens after receiving inconclusive results, we investigated whether there was any correlation between the epidemiologic characteristics of the population or the positive infection rate at that time. In this way, we hoped to learn lessons applicable to a new pandemic situation.

2. Material and methods

2.1. Study design

Initially, the correlation between the positive and inconclusive rate of SARS-CoV-2 rRT PCR over 2 years in our laboratory was investigated. This study was approved by the Institutional Review Board (approved IRB number: GCL-2022-1027, the date of approval: May 31, 2022).

During the period of the study, our lab immediately performed 2 additional retests unless all target genes were not amplified in the SARS-CoV-2 rRT PCR test. That is, when one or more target genes were amplified in the initial test, there were 3 results for each sample. And this study mainly consists of 2 parts (Fig. 1 ).

Fig. 1.

Study's flowchart.

First, on October 6, 2021–23, 2021, the results of retesting the original samples of individual SARS-CoV-2 rRT PCR tests with inconclusive results in the initial test were analyzed retrospectively. The definition of an inconclusive result of this study is the amplification of some genes among the target genes of each assay. Since the reagent used in this study had 3 target genes, when only 1 or 2 genes were amplified, it was defined as an inconclusive result. For the retesting, rRT PCR was performed by reextracting nucleic acids from the original sample, which were stored in a refrigerator at 4 to 8°C refrigerators. The retest was done twice. Allplex SARS-CoV-2 assay (Seegene Inc.; Seoul, Korea) was used for this analysis's initial test and retests. The retest was duplicated using newly extracted nucleic acids from the original sample. Data from patients previously confirmed with COVID-19 in our laboratory or possible contamination from adjacent positive specimens were excluded from the analysis.

Second, a retrospective analysis of the results of consecutive tests according to inconclusive results was conducted for 2 groups: arrivals from temporary residential facilities and locals. The Korean government operated temporary residential facilities as a space for isolation for 14 days and a SARS-CoV-2 rRT PCR test on the day of arrivals for those who entered from abroad but did not have a residence in Korea for quarantine purposes. The data of ‘arrivals from temporary residential facilities’ is an analysis of SARS-CoV-2 rRT PCR tests referred by these temporary residential facilities. 'Locals' refers to general citizens residing in Korea, and patients in high-risk facilities such as nursing homes or hospitals are excluded.

During this analysis period, all samples with positive or inconclusive results were reextracted and retested twice. The results were reported according to the retest results from January 1, 2021 to March 25, 2021 (period A) and January 21, 2022-January 30, 2022, (period B). Valid consecutive results were those of tests conducted by resampling within 7 days after the first inconclusive result. Individual and pooled SARS-CoV-2 rRT PCR test results were analyzed in this analysis.

2.2. Molecular diagnostics and interpretation

A description of the reagents used in this study is summarized in Table 1 . When a sample from a subject that has not tested positive in our laboratory in the past was requested, and if at least one of the target genes of the assay was amplified, 2 additional retests were performed in succession. For pooled tests, 2 to 5 samples were mixed to make 1 sample, and then a SARS-CoV-2 rRT PCR test was performed. When amplification of any of the target genes was confirmed, 2 SARS-CoV-2 rRT PCR tests were conducted at a time on each sample of the mixed sample [9]. The interpretation protocol for the individual and pooled test was followed in Table 2 . We used 2 automated nucleic acid extraction systems during the study period, the Dxseq viral nucleic acid isolation kit (Dxome; Gyeonggi, Korea) with the KingFisher Flex platform (Thermo Fisher Scientific; Rocklin, CA) as the main system and MagNA Pure 96 DNA and Viral NA Small Volume kit (Roche; Basel, Switzerland) on the MagNA Pure 96 platform (Roche; Basel, Switzerland) as a supplementary system.

Table 1.

The characteristics of reagents used in this study.

SARS-CoV-2 rRT PCRa	Reagent	Target gene	Cut-off cycle threshold (determined by the manufacturer)
Individual test	Allplex SARS-CoV-2 assayb (Seegene Inc.; Seoul, Korea)	E, RdRP/S, N	<40
	Real-Q Direct SARS-CoV-2 detection kit (Biosewoom Inc.; Seoul, Korea)	E, RdRP	<38
Pooled test	Standard M nCoV real-time detection kit (SD Biosensor; Suwon, Korea)	E, RdRP	<36
	Cobas SARS-CoV-2 & Influenza A/B assay (Roche Molecular Systems, Inc.; Branchburg, NJ)	E, ORF1 a/b	Not provided

^aSARS-CoV-2 rRT PCR = SARS-CoV-2 real-time reverse transcriptase polymerase chain reaction.

^bFor analysis of retest data of samples with individual SARS-CoV-2 real-time rRT PCR, only data using Allplex SARS-CoV-2 assay was used.

Table 2.

Interpretation of SARS-CoV-2 rRT PCR tests.

For individual test Amplification of target genes
Initial test	Retest 1	Retest 2	Final interpretation of the sample
Negative	No further testing is needed		Negative
All	All	All	Positive
	None	None	Negative
	Once out of 2 results, partial amplification or Ct of E gene >35		Inconclusive
Partial	All (Ct of E gene ≤ 35)	All (Ct of E gene ≤ 35)	Positive
	None	None	Negative
	Once out of 2 results, partial amplification or Ct of E gene >35		Inconclusive

For pooled test Amplification of target genes
Initial test (Pooled sample)	Retest 1 (Individual test)	Retest 2 (Individual test)	Final interpretation of the sample
Negative	No further testing is needed		Negative
All or Partial	All (Ct of E gene ≤ 35)	All (Ct of E gene ≤ 35)	Positive
	None	None	Negative
	Once out of 2 results, partial amplification or Ct of E gene >35		Inconclusive

2.3. Statistical analysis

SAS (version 9.4, SAS Institute, Inc.; Cary, NC) was used for statistics. Pearson correlation analysis was subjected to the correlation between the positive and inconclusive rates. The Chi-square test and Fisher exact test determined the difference by period and population group. A P value < 0.05 was determined to be statistically significant.

3. Results

3.1. Positive and inconclusive rates changes of COVID-19

Fig. 2 shows positive and inconclusive rates changes from February 7, 2020, to January 31, 2022. This analysis also included follow-up test results after a positive diagnosis. In the first 3 months, both the positive and inconclusive rates increased due to the local outbreak in Daegu. Pooled testing started in June 2020. Since mid-December 2020, the number of pooled tests was higher than that of individual SARS-CoV-2 rRT PCR tests, and the positive rate showed a brief decline. The r value between the positive and inconclusive rates for the past 2 years was 0.763 (P < 0.0001).

Fig. 2.

The changes in the positive and inconclusive rates with positive cases in SARS-CoV-2 rRT PCR from February 2020 to January 2022.

As the positive rate increases, the inconclusive rate also tends to increase. The r value between the positive and inconclusive rates for the past 2 years was 0.763 (P < 0.0001).

3.2. Results of retesting the original samples with inconclusive results

For this analysis, 657 results from 219 samples that had been retested were analyzed for inconclusive results. Among them, 179 samples (81.7%) showed inconclusive or weakly positive results even at retesting. According to the number of genes amplified in the first inconclusive result, there was a statistically significant difference (P < 0.0001) in the rate of inconclusive or weakly positive retest results (Fig. 3 ). Inconclusive or weakly positive results were shown in 63.1% (50/80) samples in which only 1 gene was amplified and 92.8% (129/139) of samples in which 2 genes were amplified.

Fig. 3.

Total 219 samples with initial inconclusive results (657 tests).

According to the number of genes amplified in the first test, there was a statistically significant difference (P < 0.0001) in the rate of inconclusive or weakly positive retest results. *Weak positive result was defined as a case in which all 3 target genes were amplified, but the Ct value of the E genes was over 35.

3.3. Results of consecutive tests according to inconclusive results between arrivals and locals groups

Since the overall positive and inconclusive rates for the arrivals and locals groups in periods A and B were calculated based on the number of tests, consecutive test results of subjects were included (Table 3). In both groups, the positive rate increased 6 to 7 times, and the inconclusive rate doubled in period B compared to period A. When the arrivals and the locals were compared for the same periods, the arrivals showed statistically significantly higher positive rates. The rate of inconclusive results showed a similar pattern.

Table 3.

The results of consecutive tests with new samples from subjects with inconclusive results.

	Group 1	Group 2	P-valuea (Group 1 and 2)	Group 3	Group 4	P-value (Group 3 and 4)	P-value (Group 1 and 3)	P-value (Group 2 and 4)
Epidemiologic background	Arrivals I	Arrivals II		Locals I	Locals II
Duration	Period A (January 1, 2021 – March 25, 2021)	Period B (January 21, 2022 – January 30, 2022)		Period A (January 1, 2021 – March 25, 2021)	Period B (January 21, 2022 – January 30, 2022)
Positive rateb	0.7% (237/33567)	4.6% (358/7741)	<0.0001	0.5% (4271/834993)	3.8% (10346/271325)	<0.0001	<0.0001	0.0002
Inconclusive rate	1.5% (496/33567)	3.0% (232/7741)	<0.0001	0.2% (1334/834993)	0.4% (1116/271325)	<0.0001	<0.0001	<0.0001
Number of subjectsc	301	155		707	705
Previously positive patient	0	2 (1.3%)		162 (22.9%)	28 (4.0%)
Subjects had consecutive testsd	274 (91.0%)	142 (91.6%)		197 (27.9%)	387 (54.9%)
Male (%)	244 (89.1%)	112 (78.9%)		98 (49.7%)	175 (45.2%)
Subsequent results
Positive	36 (13.1%)	40 (28.2%)	0.0002	72 (36.5%)	246 (63.6%)	<0.0001	<0.0001	<0.0001
Ct ≤ 20	2 (5.6%)	1 (2.5%)	0.4294a	27 (37.5%)	121 (49.2%)	0.0002	0.0007	<0.0001
20 < Ct ≤ 30	12 (33.3%)	9 (22.5%)		23 (31.9%)	99 (40.2%)
30 < Ct	22 (61.1%)	30 (75.0%)		22 (30.6%)	26 (10.6%)
Negative	209 (77.3%)	86 (60.6%)		101 (51.3%)	128 (33.1%)
Inconclusive	25 (10.6)	16 (11.3%)		24 (12.2%)	13 (3.3%)

^aP-value of Fisher exact test.

^bOverall, positive and inconclusive rates were calculated based on the number of tests, not the number of subjects.

^cNumber of subjects who had inconclusive results at least once during period A or B.

^dSubjects who were not previously positive patients and had consecutive COVID-19 tests within 7 days of having an inconclusive result.

A total of 1868 patients with inconclusive results and gender information were retrospectively investigated. There were some previously positive patients. In particular, it was found that 22.9% (162/707) of the locals in period A had previously positive results within 1 month.

In the arrivals, consecutive results were confirmed in more than 90% of both periods A and period B [period A, 91.0% (274/301); period B, 91.6% (142/155), respectively]. The rate of the subsequently positive patient (SPP) was higher in period B than in period A (period A, 13.1%; period B, 28.2%, respectively). For locals, consecutive results were confirmed for 197 subjects (197/707, 27.9%) in period A and 387 subjects (387/705, 54.9%) in period B. Locals also showed a higher rate of SPP in period B than in period A (period A, 36.5%; period B,63.6%).

Compared to the consecutive results of both groups by period A or B, the rate of SPP was significantly higher for locals in periods A and B. In addition, in the locals, the rate of SPP with a Ct value of 20 or less was higher than that of arrivals. For the arrivals, the rate of SPP with a Ct value of 30 or more was high in both periods A and B (period A, 61.6%; period B, 75%, respectively)

4. Discussion

In the SARS-CoV-2 rRT PCR test, an inconclusive result can lead to a lack of management in terms of prevention, and timely measures might be missed because test subjects and clinicians do not fully understand the exact meaning of the inconclusive result [10].

In this study, when inconclusive results were obtained in subjects without a confirmed history of COVID-19, the results of retesting with the original sample and consecutive tests performed with recollected samples were analyzed.

A negative result when retesting a sample with an inconclusive result implies that there might have been an error in the previous inconclusive result. This finding suggests the possibility of contamination from surrounding positive samples or previously analyzed positive samples, sample handling errors, or errors due to reagent instability [5], [6], [7]. However, only some target genes are amplified in the same way in the retest means that some target genes are actually present in the sample or that some target genes are lower than the analytical sensitivity of the reagent. Possible causes include samples from a patient within the incubation period, insufficient sample collection, and contamination during the sample collection process [11,12]. In addition, performance according to nucleic acid extraction equipment or reagent may also have an effect [13]. In particular, in the case of samples with low viral load, the difference in result values ​​depending on the extraction performance can be large, and negative results, inconclusive results, and weak positive results can be obtained from 1 sample.

In this study, 179 of 219 cases (81.7%) showed inconclusive or weakly positive results even at retest. In addition, among subjects in which 2 genes were amplified in the first result, 92.8% showed inconclusive or weakly positive results.

Therefore, in a laboratory where contamination is well controlled and testing quality is maintained, it is more efficient to report inconclusive results quickly without retesting initial samples and perform consecutive tests using new samples. This can be helpful for infectious disease prevention and control because it saves testing resources, minimizes the time waiting for results, and can reflect the confirmed result quickly. In particular, in a private referral laboratory, since it is difficult to control the sample collection process, if the possibility of laboratory contamination is low, even if the nucleic acid is newly extracted and tested, retesting using the same sample is considered to be ineffective. It is even more so when 2 of 3 target genes amplified give inconclusive results.

On the other hand, the results of consecutive tests according to inconclusive results can be interpreted differently depending on the epidemiologic background of the group to which the subject belongs and the positive rate at that time. This study compared 2 groups with different epidemiological backgrounds: the arrivals and the locals. In the arrivals, periods A and B were compared separately, assuming that the international epidemic situation would not be reflected since there was a procedure to check for negative confirmation before entering our country.

However, the results of consecutive tests of subjects with inconclusive results showed a significantly higher rate of SPP during period B. The globally confirmed cases for a week amounted to nearly 3.5 million on average in period A and over 23 million in period B. The number of confirmed cases in January 2022 was 7 times higher than in January 2021, to March 2021 [2]. Therefore, despite checking for negative confirmation before entry, the positive rate or the rate of SPP of the arrivals reflects the situation directly abroad.

Moreover, there were 13.1% of strongly positive cases with a Ct value of 20 or less in period A and 28.2% in period B. Differences in the sensitivity of the reagents used at the time of departure and entry into the country or the possibility of inaccurate sample collection at the time of testing for departure might be the cause of this discrepancy.

On the other hand, the rate of SPP was 36.5% in period A; it was 63.6% in period B in the locals. Also, the rate of SPP in the locals was higher than that of arrivals in both periods A and B. The difference in the rate of SPP according to periods A and B in the locals was statistically stronger than that of arrivals (P < 0.0001 in the locals, 0.0002 in the arrivals).

Therefore, when interpreting an inconclusive result clinically, if the subject is a local and the positive rate is high at that time, there is a high probability of receiving a positive result from a consecutive test. That is, even the same inconclusive result can be interpreted differently according to the epidemiologic background and the positive rate at that time.

In a situation where the positive rate is high, the inconclusive rate also increases, and the amount of resource input according to consecutive tests increases. In addition, the accompanying social cost also increases. Therefore, if the consecutive test according to an inconclusive result is applied more flexibly according to the subject's characteristics or the epidemiological situation, the consumption of resources can be reduced [14]. In fact, in other countries, the manufacturer has provided guidance that a result is positive even if only one of the RdRP and N genes is amplified, even though the same reagent is being used in Korea [11].

A limitation of this study is that the test's purpose, the subject's symptoms, and the history of contact with confirmed infected persons were not considered. Another limitation is that it was not possible to ascertain whether the subjects had been confirmed as positive by other institutions. In addition, we did not analyze the arrivals data according to the country of departure.

5. Conclusions

The causes of inconclusive results include several factors such as the sample; the test process, including reagents, equipment, and personnel; and the condition of the test subject. However, if contamination is well controlled in a general laboratory, retesting with the initial specimen is regarded as less helpful in terms of prevention and effectiveness versus the amount of testing resources input. In addition, inconclusive results could be interpreted differently depending on the epidemiologic background of the subject and the positive rate at that time.

Declaration of competing interest

The authors report no conflicts of interest relevant to this article.