There are multiple ways to make the diagnosis of an infectious disease. The best is growing the organism from a sterile body fluid. Although it would appear that there is no sterile body site. Bladder, lung, and perhaps even the brain have a non-pathogenic microbiome. The key word there is non-pathogenic. These organisms don’t grow in cultures. Bacteria, including pathogens, occasionally transiently access the blood and show up in cultures but can be clinically meaningless. A true positive (as opposed to a contaminant) but an unimportant positive. Still, I like a positive culture. It helps sum up my concept of ID: me find bug, me kill bug, me go home.
Next down the list of diagnostic utility is finding DNA/RNA or body parts of the organism in the blood or other body fluids. The former sometimes provides too much information and the latter can be problematic. As examples, cell-free microbial DNA assays can sometimes find DNA of multiple organisms in the blood, none of which were clinically expected or relevant. And the 1-3 beta-D-glucan, which detects a fungal body part, may be elevated when there is no fungal infection. Like all tests, they have to be judged within the context of the clinical presentation.
The most problematic way to make a diagnosis is serologies: measuring antibodies in the blood directed against an infection.
There are two antibodies that can be used to diagnose an infection. IgM, which is the first antibody class the immune system makes against an infection. It kicks in around day 5-10, peaks around 21 days and is present in the blood for around 3 months. Usually. There can be great variability depending on the infection and IgM can be problematic.
IgM tests also suffer disproportionately from false-positive results, which can lead to misdiagnoses, inappropriate therapy, and premature closure of a diagnostic workup.
IgG is the other useful antibody. It kicks in around day 5-10 and lasts a lifetime. IgG is the antibody that can give lifelong protection and the antibody whose levels can be boosted with vaccines or re-exposure to the pathogen. It is the only wing of the immune system that can be boosted. IgA and IgE have no routine diagnostic utility.
If you have a positive serology, then you must have the disease. Right? I wish. There are several serologic scenarios that drive me bonkers. The first is Epstein Barr Virus (EBV).
It is not uncommon for EBV serologies to be ordered in clinical situations that I call Nutrimatic medicine: a clinical presentation that is almost, but not quite, entirely unlike EBV. You can fill in other pathogens for EBV. But it is remarkable how often tests are ordered when there is no clinical indication. Diseases have patterns. If the pattern is not that of, say, EBV, then do not order EBV serologies.
Often EBV serologies are ordered in patients who have a fever and only a fever. None of the clinical features of EBV like a sore throat, enlarged lymph nodes, hepatitis, cytopenias etc. And the test is positive. Then the patient is sent to me to explain the serologies.
Since most people are exposed to EBV when young, the majority of the population will have positive serologies. To complicate the issues, EBV, like all herpesvirus, persists in human cells. Unlike love, herpesviruses last a lifetime. When a process activates the immune system, it can also activate herpesvirus replication, leading to a boost and higher titers of antibodies.
I remember back when I was an intern, EBV, because of high EBV titers in Chronic Fatigue Syndrome (CFS) patients, was thought to cause chronic fatigue syndrome, rather than a marker of CFS, and not a good marker at that.
I would like to say those who order EBV serologies are required to interpret that serology, but so often the person ordering the tests is, well, let’s say lacking in a sophisticated understanding of EBV.
So I have to tell the patient: the serologies is meaningless since (almost) everyone has positive serologies, the test should not have been ordered in the first place, and EBV has nothing to so with your, usually resolved, symptoms. Blah.
Lyme testing is the other problematic serology. Lyme testing has two steps. The first is a screening ELISA, meant to be overly sensitive. Then a confirmatory western blot. There are many other Lyme tests, such as PCRs, antigen tests, and labs that offer their own, in-house, tests. All of these are problematic.
Lyme testing gets to the heart of the matter of prior plausibility in interpreting the results of serologic tests.
There is zero Borrellia in the Willamette Valley. None. Zilch. Nada. At least as of now. But who knows with the future. One of the many consequences of climate change is the range of infection vectors is changing. The ticks that spread Lyme, and other diseases, are spreading north and east, so perhaps Borrellia will migrate to the Great Pacific Northwest some day. And pathogens can spread due to human behaviors, such as Rocky Mountain Spotted fever in Arizona due to pet dogs. Pathogens get around, sometimes by the damndest methods. So you can never be 100% sure there wasn’t some weird method of infection acquisition.
Every test has a false positive and false negative rate. It is one of the annoying characteristics about serologies. The results are not black and white, but shades of probable gray. I sometimes wonder about the sensitivity of some tests. For example, the urine antigen test used to diagnose histoplasmosis was developed from a strain out of Indiana. Strains in the Caribbean, which somewhat antigenically different, can be missed with this test. For some organisms, the farther you are away from the index strain, the more genetically separate they become. It can be quite dramatic with malaria, where the P. ovale in Asia look and acts identical to the P. ovale in Africa, but the two cannot mate. Classic Darwin.
That may be true for Lyme as well, as the farther you get from the index Borrellia of the NE, the more antigenic distance you see. Does this make a difference in testing? Don’t know for strains in the US, but the US testing for Lyme is not as accurate for detecting European Borrellia.
And then there is the potential of cross reacting with other spirochetes, like syphilis and relapsing fever.
Lyme serology, like other serologies, is not as simple as one would like. And it is compounded by the fact that, like EBV serologies, they are ordered when they shouldn’t be.
For example.
The patient lives and works in suburban Portland and has not left the area for a couple of years. He does find a tick after yard work which he notices immediately and removes the tick. It happened to me this summer as well. I found two attached ticks on my abdomen after a day trimming bushes. It gives me the willies. Ticks and leeches. Ew. Second most horrifying scenes in the movies is when Bogart has to remove all those leeches in the African Queen. The worst? Maybe because I was just a kid when I first saw it, but it was the amputation scene in Gone with the Wind. I digress.
As mentioned, there is no Lyme in Oregon. And the tick has to feed for at least 4 hours before passing on Lyme. So no reason whatsoever to test for Lyme. But the patient asks his primary physician for testing and money is wasted when it come back negative. At least the patient did not see an ND (Not a Doctor), who love to send off non-standard Lyme testing that often comes back positive. But that is a story for another day.
But the patient has vague orthopedic complaints. Muscle and joint aches that are new in his 50s. Just wait until you hit the 60s. I do not think I have a single part of my body that doesn’t ache. Maybe my hair. And that’s likely because I do not have much. Six months in, because of the symptoms, the patient requests further testing, so a very-much-not-indicated western blot is ordered. Really. Might as well just burn a stack of dollar bills.
The western blot was negative for the IgG and positive for IgM. He has Lyme!
Nurp.
First, if this were Lyme, which it isn’t, the blot would be consistent with acute Lyme. Six months in the IgG should be positive as well. It isn’t. And the IgM western blot is notable for false positives.
But more importantly, a positive test done in a patient with no risks is going to be a false positive. That’s the SBM part of this entry.
In trying to figure out what a Lyme, or other serology, means, you have to know not only the sensitivity, specificity, false positive and false negative rates of the test, you also have to know the prevalence of the disease. And with Lyme, the added complexity of the operational characteristics of the test also depends on the stage of disease.
So how to put that information to work? Wiser heads than I have discussed this, for example here and here. I am lazy and the statistics, especially Bayes, makes my brain hurt. But you can plug the information into the formula and do the calculations yourself. It is the concept I want to emphasize.
When the incidence of a disease is low, most serologic positives are false positives.
As the incidence of disease falls in a community, the more likely a positive serology is a false positive and the patient does not have the disease. But it is simpler in Oregon: no Lyme in Oregon equals no Lyme in the patient. So I explained to the patient why the result was meaningless and he should never have asked his doctor for the Lyme testing. And the doctor should not have ordered the test. When I get these requests, I tell the patient I am not a drive up the window at McDonald’s. Unless they want an Oreo McFlurry.
This is contrasted with the patient I saw the same week who, after a tick bite in Connecticut (the state, not a nickname for a body part) three months ago, now has a Bell’s palsy. I almost didn’t need Lyme testing to know what the patient had.
Not all serologies are as problematic as Lyme, but it is remarkable how often I see serologies ordered in patients who have almost zero risk for the disease. If I were a patient, I would be pissed if part of my copay was for completely unneeded tests.
A couple of years ago I was admitted for less than 24 hours for small bowel obstruction. My bill was $10,000, I had to write a check for 10%. I looked over my charges, and there was nothing on the list I considered inappropriate. But if I had to pay for even 10 bucks for an unneeded Lyme test, I would be pissed.
Ordering these tests a waste of money and time that often only adds confusion to the case and an unneeded ID consult. I wonder about the ‘unneeded’. Some days I think all ID diagnostics and therapeutics should require ID prior authorization. I have this recurring nightmare where oncology is approached like ID. Multiple fear driven therapies are given in the absence of appropriate diagnostics.
Lyme and EBV are not the only serologies that gives me a headache. Syphilis? That would be a blog entry in itself to go through all the issues with that infection.
The take home: A positive serology for an infectious disease may be meaningless.