There are a variety of tests that can help diagnose Lyme disease by finding evidence of Lyme infection. These include ELISA, IFA, PCR, western blot, elispot, and blood culture. Although these tests are helpful aids they are far from perfect. In this article I review each of these tests and give them a grade. With these grades in mind, I describe the way I use these tests to help decide if someone has Lyme disease.
(Note: I do not include the CD-57 test in this review because it only shows immune suppression which may or may not be caused by a Lyme infection. For my review of this test see CD-57 Test? Rarely.)
As you read about the tests, note that a test does not diagnose a disease. As a physician, I do not treat a test; I treat a person. Tests can be wrong. To make a diagnosis I consider
- the risks of getting the infection like having a known tick bite, or a history of hiking where there are a lot of ticks that carry Lyme like in Minnesota, or a number of other risk factors,
- the symptoms,
- physical exam findings, and
- whether there is supportive testing.
Note that I said supportive testing. So the result of testing is only one part I consider in making a diagnosis. See How to Diagnose Chronic Lyme Disease for more information about this complex issue.
Terms and Definitions
To help understand when to use a test or the meaning of a result physicians consider the test sensitivity, specificity, prevalence, and predictive value.
Sensitivity is the ability of a test to find an illness in all people with the illness.
Specificity is the ability of a test to correctly identify people without an illness from all people who do not have the illness.
Sensitivity and specificity are related. In general if a test is very sensitive and can find a very high percentage of people with an infection it is usually less specific. Sensitivity is the ability to find an infection when present while specificity indicates the accuracy of the positive test. A highly sensitive test often has many false positives and a lower specificity.
Let me use a fishing analogy. To catch 100% of salmon in a one mile square section of the ocean using a net a fishing boat will also capture many mackerel, halibut and other fish. That is to say it is very sensitive for finding all of the salmon. However fishing for all of those salmon using a net is not that specific because many other types of fish are also pulled in.
Prevalence is the amount of an illness as percentage or proportion of certain population.
One type of population is a geographic area like the state of New Jersey or my home state of Indiana. The tick population of New Jersey carry a lot of Lyme germs compared to Indiana ticks. Because of this Lyme disease is much more prevalent in the people of New Jersey than in the Hoosiers of Indiana.
Here is another example. In people who have a lot of Lyme disease symptoms there are many more who have the Lyme infection compared to people who do not have Lyme disease symptoms. That means Lyme infection is more prevalent in people with many symptoms of the illness compared to those who have no or a small number of symptoms.
Predictive value is the chance a test result is correct. In other words it indicates if a positive result is really positive or a false positive. Is the result of the test real? The predictive value depends on the sensitivity and specificity of a test as well as the prevalence of an illness.
As an example when a test for Lyme infection is performed on a person from New Jersey who has a lot of Lyme disease symptoms a positive test has a very high predictive value. When a test is positive it has a low chance of being a false positive. By comparison a positive test for Lyme infection from a Hoosier in Indiana who does not have many symptoms of Lyme has a very low predictive value. It means a positive test is likely a false positive.
The Tests and Grades
I. Antibody Tests
Antibodies are proteins the immune system develops to fight infections. Antibodies attach to germs. Once a germ has antibodies attached to it immune system cells can identify the germ and kill it. Essentially antibodies mark a germ for destruction.
ELISA and IFA: Grade F
These are screening tests to see if a person has antibodies that attach to the covering of Lyme germs. Unfortunately not only do Lyme antibodies attach but other types of antibodies made against viral infections or other germs can also attach. Because of this these are “could” tests. When these tests are positive they indicate that a person could have antibodies against a Lyme germ.
As a screening test for Lyme the ELISA and IFA are very poor. In fact numerous studies indicate an ELISA test only has a sensitivity of 50%.
Believe it or not, even with such poor reliability, the CDC and the Infectious Disease Society of America (IDSA) recommend this as the first test to perform on a person suspected of having Lyme disease. They recommend a two step method that includes an ELISA or IFA as step one. If one of these are positive then the second step is to perform a western blot which I describe below.
Grade: F. These tests have such a low sensitivity that they should not be used. They are an absolute waste of money even when performed by a high quality Lyme testing lab such as IGenex.
Western Blot: Grade B
A western blot is a method to detect Lyme antibodies. It identifies antibodies that attach to very specific proteins found on the covering a Lyme germ. The proteins are identified based on what they weigh and are called bands. They are measured in kilodaltons (Kda) which is a microscopic unit of weight. There are 6 bands with antibodies attached that can indicate a Lyme infection. These are the 18 Kda, 23-25 Kda, 31 Kda, 34 Kda, 39 Kda, and 83-93 Kda proteins. Some physicians do not consider the 18 Kda to be one of these. To increase the predictive value of the test for a person, it is best to have antibodies against at least 2 of these bands. Having 2 bands present though decreases the test sensitivity.
IGenex at one time included the results of its own validation studies on its western blot reports. In these studies IGenex required two positive bands from the following 23-25, 31, 34, 39, 41, and 83-93. The tests were performed on blood samples known to be positive for Lyme 36% of the time. Even though band 41 is not that specific for Lyme IGenex still includes it in determining a positive Lyme test. Using the IGenex method and interpretation standards the sensitivity and specificity of an IGM antibody western blot is 73% and 96% respectively and an IGG antibody western blot is 63% and 96%.
By comparison the western blot interpretation methods approved by the CDC/ IDSA on the same sample of blood found an IGM sensitivity of 58% and a specificity of 99%. For the IGG the sensitivity was only 38% and the specificity was 100%.
The sensitivity of the IGenex interpretation standard shows it is a better test at finding Lyme than the CDC/IDSA method, but is less accurate when positive with a lower specificity. In reality though both tests have low sensitivity and will miss many with Lyme.
To increase the sensitivity of the tests many physicians such as myself will ignore the IGenex and CDC/IDSA standards and interpret a test as positive when only one or more of the following bands are positive 18, 23-25, 31, 34, 39, and 83-93. Many of us also will consider indeterminate findings at one of these bands as weakly positive.
Grade: B. The lower sensitivity of this test lowers its grade from an A. However the sensitivity is increased using the increased sensitivity interpretation standards I use noted above. And yes this could increase false positives, but this test is considered along side a patient’s symptoms, physical findings, and risk of getting the illness in diagnosing Lyme disease.
One more thing, there are newer tests like the Lyme blood culture and the iSpot tests I describe below. The western blot test has a higher grade than these because I and others have a longer history using them and know how it behaves in the real world.
PCR: Grade C
A polymerase chain reaction (PCR) tests finds the genetic material (DNA) of the Lyme germ (borrelia) in whole blood. In a PCR test the DNA of a germ found in blood is amplified so that it can be seen. The test however has a sensitivity of nearly 30% when all studies are considered. Thus it is not that useful at finding a Lyme infection. However, when it is positive, the test really is most likely a true positive.
There are different reasons the test has such low sensitivity. The main one is Lyme germs rarely live in the blood. Oxygen in blood is toxic to borrelia. Borrelia is a real trickster and master of survival. It tends to live in tissues that do not have much blood flow like in tendons or on muscle fibrous coverings. So finding the Lyme germ and its DNA in the blood is like finding a needle in a hay stack.
Grade: C. PCR has a very low sensitivity and really is not good for finding Lyme. The only reason I do not give it a D or lower is that when the test is positive it really is positive.
Elispot (iSpot): Grade C-
An elispot test identifies if immune system cells called T cells have been exposed to Lyme germs. In the United States the test is a newer test available as an iSpot through a lab called Pharmasan labs. It has been available in Europe for sometime currently through InfectoLab.
In an elispot T cells from a person with Lyme are exposed to certain antigens (protien complexes) found on the covering of a Lyme germ. When T cells are exposed to these antigens they release chemicals including gamma interferon. T cells in the test will only release gamma interferon if they have been exposed previously to the Lyme infection in the person being tested. There are false positive tests due to other causes of inflammation. Pharmasan has a technique to to take this into account before it reports a positive test.
The one advantage to this test is T cells become reactive to Lyme in about 2 weeks after an infection. The western blot may not turn positive for 4 or more weeks.
There is only one study that has been done looking at the iSpot. And it was performed by Pharmasan. So there is a bias for Pharmasan to find good results about its test. In the study 500 people were tested. People diagnosed with Lyme disease were in one group and people without Lyme were in another group (a control) group. In the study the iSpot had 84% sensitivity and 94% specificity. In the publication of the report there is not an indication of how many of the 500 people were in the control or disease groups. In addition, the standards used to determine those that had Lyme are not stated. In a recent call I asked one of the researchers about the standards to determine who was put into the Lyme disease group during the study and he noted those who were diagnosed with Lyme disease by 2 or more physicians who did not use any agreed upon standards.
The reason I raise the issues about the test design and reporting is that it cannot be replicated or verified based on other researchers. So it is hard to say if this test is really valid or accurate based on other research. In addition only 500 people were in the study, so it was too small of a study to determine if the results have any meaning in the real world where I see patients. By comparison the western blot has been studied by many. And I and other physicians have extensive experience using and interpreting it so we know how it behaves in the real world.
Grade: C-. Although this test has good sensitivity and specificity in one study, it is hard to say if in day to day use it will prove to be effective. However, it does provide another measurement of immune cell activity not addressed in a western blot. In addition, it can find Lyme infection earlier than a western blot.
One other reason I do not give it a higher result is insurance considers it experimental. Because of this insurance will not pay for the test. In addition insurance companies will not recognize it as a test that proves Lyme infection when making determinations about whether to cover IV antibiotics or other Lyme treatments. So I do not think it is worth the cost.
Lyme (borellia) Blood Culture: C-
In the last two years a United States lab called Advanced Laboratory Services (ALS) has developed a Lyme disease blood culture that may have good predictive value in finding Lyme infection. Traditionally the problem with Lyme blood culture has been finding a substance to grow the Lyme germ. ALS has perfected a method that involves three separate steps. In the first step the blood is incubated on a special modified BSK medium. Then after 2 weeks it is transfered to a medium that has a protein matrix. As a third step a PCR techinique is use to determine if the growth on the medium is really borrelia. PCR is a techinique that finds the dna of an infection, in this case borrelia.
There is one study of this technique performed by Eva Sapi PhD and her research group. Of note the research was supported by ALS and Dr. Sapi is a member of the Research Division at ALS. So again like in the iSpot test above, this means there is an inherent bias built into the results of the research to find a positive result supporting the effectiveness of the test.
The test included 72 people who were known to have Lyme based on the very strict standards of the CDC and IDSA. There were also 50 people put in a control group who did not have Lyme. The study results found 94% sensitivity and 100% specificity.
Perhaps the test was very sensitive in the study because the people with Lyme enrolled in the study met the strict CDC/IDSA diagnostic standards. It is possible those meeting this standard have more Lyme living in their blood than the average patient I treat in my office who does not meet the CDC/IDSA standard. In addition, the study sample was too small to determine how accurate the test is in the real world.
In fact in my limited use of this test I find it misses many that I diagnose with Lyme disease based on other tests and clinical grounds.
Grade: C- Although this test based on one research study has good sensitivity and specificity, I doubt it is accurate in the real world. The main reason for this is that Lyme does not readily live in the blood, and I believe is less likely to do so in those that do not meet the strict CDC/IDSA diagnostic criteria like those enrolled in the Sapi study. I suspect the real world sensitivity of the ALS blood culture is higher than the 30% sensitivity of a PCR test but much lower than the 94% found in the Sapi study. As I point out in the PCR section above, the major reason PCR sensitivity is low is that Lyme does not live in the blood. The Lyme blood culture is designed to overcome some of this problem by amplifying even one germ found through the culture technique, but the problem remains.
Having said this, I applaud Dr. Sapi and her advisors Alan MacDonald MD and Joseph Burascanno MD for finding an effective media to grow borrelia. This is a real milestone and far surpasses the effectiveness of previous culture techniques. I am interested in learning about the usefulness of this method in my real world practice here in Seattle, WA USA where I treat people from all over the world. I may be wrong in my assessment, but time will tell. I am open to learning otherwise.
One other reason I do not give it a higher result is insurance considers it experimental. Because of this insurance will not pay for the test. In addition insurance companies will not recognize it as a test that proves Lyme infection when making determinations about whether to cover IV antibiotics or other Lyme treaments. So I do not think it is worth the cost. And for the record at the time of writing this article the cost of the ALS blood culture for borellia is $595.00 US dollars. By comparison the cost of doing a western blot for IGM and IGG antibodies through IGenex is $200 US dollars.
When and What Test
So this is a lot of information to digest and to figure out how to use. Here is what I recommend and do in my practice.
- Never do an IFA or an ELISA.
- The first and only test I recommend at an initial visit for someone who has not been diagnosed with Lyme disease is the Western Blot. Be sure to have it performed by a lab that tests and reports for all of the Lyme specific bands including 18, 23-25, 31, 34, 39, and 83-93. These labs include IGenex and the Lyme Disease Labarotory at Stony Brook. Note that the western blot in Canada does not test for all of these. In addition all commercial labs like LabCorp, Quest Diagnostics, PAC LAB, and PAML here in the United States use the FDA approved kit using the CDC/IDSA criteria and do not test for all of these bands either.
- If a western blot is negative and it is not clear based on clinical grounds that someone has Lyme disease then perform either the iSpot (elispot), PCR, and/or ALS Lyme blood culture. My preference here still is the PCR because it is an older test, and I have more experience working with it.
- I am ok treating people without a positive test based on clinical grounds. However, treatment can be difficult and can sometimes take months to see improvements. For this reason some of my patients want to have a positive Lyme test because it is easier to continue treating if a test is positive. If this is so, I will perform any of the tests in 3. above one at a time or all at once. To read more about how long it takes to improve and recover from Lyme see Getting Healthy. Will I? When? How do I?.
- I do not recommend Lyme testing panels. Some labs have panels that include PCR, IFA or ELISA, western blot testing for Lyme with or without a CD-57 test. In my opinion these panels are only good for the lab which profits from unnecessary testing.
I opened this article stating that Lyme is always a clinical diagnosis. Again a test does not diagnose a disease. Also be aware based on the concepts of predictive value and prevalence I present above. Based on these concepts there is a great chance of a false positive test in someone who does not have Lyme disease symptoms. For more information see How to Diagnose Chronic Lyme Disease.
An Understanding of Laboratory Testing for Lyme Disease. Harris et al. Journal of Spirochetal and Tickborne Diseases—Volume 5, Spring/Summer 1998.
iSpot LymeTM: A New Generation of Lyme Disease Testing. Chenggang Jin, PhD, MDa,b, Deanna J. Fall, BAc, Diana Roen BS, and Gottfried Kellermann, PhD c. ©NeuroScience, Inc. April 2013. (You can find a copy on the Neuroscience website.)
Improved Culture Conditions for the Growth and Detection of Borrelia from Human Serum. Eva Sapi et al. International Journal of Medical Sciences 2013; 10(4):362-376.
by Marty Ross MD