Informing the public involves clearly acknowledging what is still not known about this virus, and it involves stating what tests simply can or cannot do
By Dr. Edsel Maurice T. Salvana, MD, DTM&H, FPCP, FIDSA
Mass testing or targeted testing? RT-PCR, rapid antibody test, or antigen test? Which test is best to diagnose Covid-19? Which test is best to rule in or rule out infection? These questions have been the subject of intense, sometimes acrimonious, debate as the world struggles to control the pandemic.
Let’s start with understanding what medical tests do, and how to interpret them. There are two reasons to test, for clinical diagnosis and for public health surveillance. Clinical testing is undertaken to help a doctor decide on what to do with a patient. Surveillance testing is used to confirm the diagnosis and to keep track of a disease’s prevalence in a community. In the case of infectious diseases, surveillance testing tracks the spread of the disease. This is important in outbreak situations.
Clinical testing is warranted when the doctor needs to decide how to proceed after a clinical assessment. A clinical assessment is made after a thorough history and physical examination. A “pretest probability” guides the doctor’s decision to use a test if it is needed. Pretest probability is the likelihood that a person has the disease in question before the test is done. This is determined by a trained medical professional, based on clinical assessment in combination with knowledge of the disease process.
For instance, a patient in the Philippines with a fever and a low platelet count, living in a community with an ongoing dengue outbreak, probably has dengue. He would have a high pretest probability, and a dengue test may not be necessary because a negative test would not change your plan to treat the patient for dengue. If a similar patient had a fever but his platelet count was normal, a dengue test would help make the diagnosis. In the second case, it is possible the patient has dengue but there aren’t enough clinical features to say for sure. On the other hand, a similar patient in the US with no history of travel to the Philippines would probably not be tested for dengue, but would undergo a workup for other diseases such as idiopathic thrombocytopenic purpura.
Surveillance testing is used by public health authorities to determine the behavior of a disease and track its spread. While it may not be necessary from a clinical standpoint to test the suspected dengue patient with low platelets and fever to guide management, documenting the infection can help authorities decide on public health interventions such as destroying mosquito breeding areas. This can guide policies that would prevent emergency rooms from overflowing with dengue patients who would receive substandard care in an overwhelmed healthcare system.
There is no 100 percent accurate test. Test accuracy is usually measured using sensitivity and specificity. Sensitivity is the likelihood that a test is positive when the disease is present. In other words, if you are sick, sensitivity is the probability that the test will come back positive. Specificity is the likelihood that a test is negative when disease is not present. Specificity is the probability that if you are not sick, the test will come back negative and disease is effectively ruled out. Testing can be very confusing even for trained professionals. It is easy to misuse a test if you are not familiar with the test characteristics and the disease process.
Some tests are best used for people with active disease, while some tests are best for diagnosing past infection. In the case of Covid-19, RT-PCR and antigen tests can detect active infection, while antibody tests can pick up past infection. Each test has its strengths and weaknesses, and different kits have different performance characteristics. A test is just a snapshot in time. You can test negative one moment and get the disease when you walk out the laboratory, so it needs to be treated as true for only a short period of time.
RT-PCR stands for reverse transcription polymerase chain reaction. Reverse transcription is the process of turning viral genetic material from RNA into DNA. Polymerase chain reaction works like a photocopying machine that replicates minute traces of reverse transcribed viral RNA. While RT-PCR is very sensitive when performed on the correct specimen, the method of collection can affect the overall sensitivity. Nasopharyngeal swabs are on average about 70 percent sensitive and nearly 100 percent specific if properly collected. RT-PCR is considered the best test for the detection of SARS-CoV-2, the virus that causes Covid-19. RT-PCR becomes positive after the third day of infection. It takes an average of five days from the time of infection to manifest symptoms. RT-PCR can detect asymptomatic carriers about two days prior to the onset of symptoms. If it is done on the first three days from exposure, it will most likely be a false negative because there isn’t enough virus present yet. The best time to do RT-PCR is on the third day of symptoms (eighth day of infection) when the sensitivity reaches 80 percent.
RT-PCR doesn’t distinguish between infectious virus versus viral remnants that can persist for up to eight weeks after recovery. This is why patients with asymptomatic, mild, or moderate Covid-19 are no longer retested after 14 days. There is typically no more infectious virus present after two weeks of illness even if the RT-PCR remains positive. RT-PCR is expensive and highly technical, and the turnaround time is about two to three days.
Antibody testing can take the form of a rapid antibody test, an ELISA (enzyme-linked immunosorbent assay) or a CLIA (chemiluminescent immunoassay). A blood sample is needed for antibody testing. The rapid antibody tests have widely varying sensitivities and specificities and will not pick up disease reliably before 14 days from the start of symptoms. DOH and WHO do not recommend using an antibody test for diagnosis of active disease since it is not reliable as a standalone test. It also cannot reliably predict immunity, and it might not become positive at all in mild disease. A positive antibody test can become negative after a few weeks. Antibody tests, especially ELISA and CLIA, can be useful for retrospective diagnosis during serologic surveys and for measuring the burden of illness in a community. These are usually done for public health surveillance.
Antigen tests are the new kids on the block. These look for a piece of the virus and are useful for diagnosing acute or ongoing disease. A nasopharyngeal swab is used for this test. These tests are less sensitive than RT-PCR since they require high viral loads to turn positive. They are, however, much easier, cheaper, and faster to perform. Turnaround time is only 15 to 30 minutes, and some antigen tests do not require any specialized equipment. Specimens still need to be collected carefully with full PPE and infectious waste needs to be disposed of properly. The main advantages of antigen tests are speed and the ability to detect people who are actively infectious. A positive test need not be confirmed with an RT-PCR since antigen tests are highly specific. A negative test may need to be checked with an RT-PCR depending on the pretest probability. A repeat antigen test may also be done when in doubt.
So who needs a test? It depends. Someone from Batanes probably won’t, since there are no documented cases of Covid-19 there. Someone from Metro Manila who has symptoms probably needs a test. What about an asymptomatic factory worker who lives in an area with community transmission? This is a more nuanced problem. A recent paper showed that symptomatic people are more likely to transmit Covid-19, but there remains a small risk that an asymptomatic infected person can do it. Will any of the tests be able to detect these asymptomatic spreaders at all? The most effective strategy is still to Covid-proof the workplace with masks, physical distancing, face shields, and good ventilation, since it may be almost impossible to pick up all these hidden spreaders. Whether or not mass testing, serial testing, combination testing, or no testing can mitigate spread remains an open scientific question. Ultimately, any testing decision needs to be made by doctors who understand how to use these tests and what the results mean.