Covid Tests: Investigating Errors

News circulated in the social media about the discrepancy in COVID-19 test results from 2 reputed laboratories of Valley has become a matter of concern. The swab samples were taken on 2 different days. The lack of reproducibility of test results could be due to sampling or to rapid fall in the viral load over the interim short period.  Here, we comment on COVID-19 testing and critically analyze potential errors which may affect the outcome of testing. We have not inspected or reviewed the working of either laboratories and the document has no reference to working of these laboratories. We have no conflict of interest in this writing except for service to the society in disseminating knowledge and allay apprehensions.

Testing for coronavirus infection has become the cornerstone to fight against the COVID-19 pandemic. World Health Organization Director-General said at the March 16, 2020 media briefing on COVID-19, “You cannot fight a fire blindfolded. And we cannot stop this pandemic if we don’t know who is infected. We have a simple message for all countries: test, test, test. Test every suspected case”.  One of the most important tools to control COVID-19 spread namely ‘contact tracing and isolation’ depends on testing and identifying infected persons. Countries like South Korea with their policy of ‘drive through testing’ have successfully reduced the infection rates. There are large increases in the numbers of COVID-19 testing centers and an increasing number of people are being tested in every country.  By June 1, 2020,   681 laboratories (476 governmental and 205 private) have been set up in India, with a daily capacity of around 1.2 lac tests. The process of setting up these centers and quality control has been under the direct supervision of the Indian Council of Medical Research (ICMR). It is commendable that many such laboratories have been set up in places where there was little knowhow or expertise related to nucleic acid testing, the basis for diagnosing COVID-19.

The gold-standard for diagnosing COVID-19 infection depends on a technique called ‘Real-time reverse-transcriptase polymerase chain reaction’ (Real-time RT-PCR). The test detects the ribonucleic acid (RNA) of SARS-CoV-2, the causative agent of COVID-19. The test has three stages: (1)  sample collection, (2) RNA extraction, (3) polymerase chain reaction (PCR). A nasopharyngeal or oropharyngeal sample is collected by a swab into a sterile tube containing Viral Transport Medium. For nasopharyngeal sampling, the swab is inserted in the nostril and gently moved forward into the nasopharynx and rotated for a specified period to collect secretions that contain the virus in infected subjects. The extraction of the viral RNA from the collected material is done by using a lysis buffer and silica-based cartridges.  The genetic material of coronavirus is a single-stranded RNA of 30kb. To detect it, the RNA must be converted into complementary DNA (cDNA) sequences by reverse transcriptase, and the newly synthesized cDNA is amplified into millions of copies by standard PCR procedure (RT-PCR). In the current test, three separate regions of the viral RNA are amplified to give better sensitivity and specificity. These are detected with specially labeled probes designed to bind within the amplified regions. As the viral genetic material is amplified and its concentration increases in the tube, the binding of the probe gives an increased fluorescent signal that can be detected in real-time. Thus, the technique allows seeing the results almost immediately while the process is still ongoing (Real-time), whereas conventional RT PCR only provides results at the end of the process.

The basic laboratory equipment needed for the test is a thermal cycler (also known as a thermocycler, PCR machine, or DNA amplifier) to amplify segments of cDNA via the polymerase chain reaction (PCR). A real-time PCR machine also can continuously detect fluorescent signals. Several reagents are needed for the test which includes lysis buffer, Reverse transcriptase and DNA polymerase enzymes, nucleotides, forward and reverse primer sets, and probes.  These are supplied as kits that are available from several suppliers. ICMR has validated over 40 kits based on its performance for COVID-19 testing. Manufacturers evaluate the kits for precision (can detect around 10 viral copies/ul) and reproducibility on repeated tests, different instruments, and by different operators. The diagnostic accuracy of all the validated kits on confirmed positive and negative samples is over 99%. Based on these findings, Real-time RT-PCR is considered the gold standard of diagnosis for SARS-CoV-2 and many other viruses, with high sensitivity and specificity in a laboratory setting.

While Real-time RT-PCR is a highly sensitive test for the detection of extremely minute quantities of viral genetic material, there are several inherent problems when these tests are used in the clinical setting. Both false positive and false negative results can occur with undesirable consequences. COVID-19 false-positive test can put the individuals and their family into unnecessary hardships of the lockdown and associated social stigma.  A false-negative test can risk the fight against COVID-19, as the infected individual can continue to transmit the infection without being isolated. So, it is important to critically evaluate what are those situations which can introduce errors in testing and be aware of those while performing the COVID-19 Real-time RT-PCR tests.

The most important variable in Real-time RT-PCR testing for COVID-19 is the sampling technique and the site and time of sample collection during the disease course. Most experts believe improper sample collection, exemplified by failure to acquire enough sample and failure to insert a swab deep into the nose, are to blame for the low clinical sensitivity and leading to false-negative results. Also, improper sample handling and/or storage can cause sample disintegration and resultant false-negative results. The site of a collection of swab samples has a major impact on test accuracy. The positivity rates for samples from different sources vary widely: bronchoalveolar lavage fluid 93%, fibro-bronchoscopy brush biopsy 46%, sputum 72%, nasal swabs 63%, pharyngeal swabs 32%, feces 29%, blood 1%, and urine 0%. Thus, nasal swabs, the gold standard for COVID-19 testing can detect only 2/3rd cases and pharyngeal swabs will detect only 1/3rd cases. Next, the likelihood of detecting the virus genome depends on how much time has passed since infection. In one study Real-time RT-PCR was positive in 67% in the first week after onset of symptoms, 54% in the second week, and 45% in the third week. In contrast, coronavirus antibodies were detected in 93.1%, 82.7%, and 64.7%, in first, second, and the third week of disease respectively.  Combining RNA and antibody detections significantly improved the sensitivity of pathogenic diagnosis for COVID-19.

The second problem which can cause errors in Real-time RT-PCR is contamination during testing. Contamination is likely to occur if the standard operating procedure (SOP) is not adhered to during testing. The choice between one-tube versus two-tube RT-PCR has important consequences. In the one-tube procedure, being followed for COVID-19 testing, the reagents for both 1st-strand cDNA synthesis and PCR amplification are combined into a single tube.  Complementary DNA synthesis and PCR take place successively in a specially optimized RT-PCR buffer during an uninterrupted thermal-cycling program. Because the reaction tube is not handled between cDNA synthesis and PCR amplification, the one-tube RT-PCR system requires less hands-on time and involves a lower risk of sample contamination than two-tube systems. However, contamination is yet possible as the system is not completely closed. Unsuitable handling of samples containing a high concentration of COVID-19 viral RNA or positive control template can generate significant viral aerosol and contaminate an otherwise negative sample between RNA extraction and RNA amplification. This can lead to false-positive results.

Lastly, we have to be cognizant of the unique circumstances of COVID-19 testing. The expansion of testing facilities has occurred in an extremely difficult and life-threatening situation. Some laboratories have been set up in places where nucleic acid testing was never done earlier. Technical staff, not trained in such technology have been trained in the shortest possible time, with a limited learning curve. On top of that, each laboratory is loaded with hundreds of samples to be done in a setting that is time-consuming and not entirely automatic. The technical staff is geared up to protect themselves, making operability difficult.

With all this, we feel proud that hundreds of samples are being tested on daily basis in many laboratories across the country. However, all this puts up many issues and unintentional errors in the system. We believe that the efficiency of any procedure especially highly sophisticated ones like Real-time RT-PCR for COVID-19 will need a proper learning curve, which occurs over time. The test performance will get better with time as the teams gain learning experience and confidence. It is commonly observed that a third-year resident doctor performs better than a first-year resident and the same is true for laboratory work, including the use of real-time RT-PCR.

With all this, we are proud of these teams who have performed so much and so well under extremely difficult and tiring times. God bless all these COVID-19 warriors.

Prof. M. S. Khuroo, former Director, Sher-I-Kashmir Institute of Medical Sciences, Srinagar, Kashmir.

Dr. Shahid Jameel, Ph.D., CEO Wellcome Trust DBT India Alliance.