Many people notice a single lab value—a flagged result, a lower eGFR, or a higher creatinine—and immediately become alarmed. In this talk, I explain why drawing conclusions from a single snapshot can be misleading, how to properly interpret what eGFR represents, and what truly points to kidney damage. We also discuss what can temporarily influence lab results and why kidney function can only be assessed responsibly by looking at trends and patterns across multiple measurements, rather than in isolation.
The content of the video in written form
Hello, this is Dr. Zsom.
Dear patients, well, in science and among the sciences, in medical science, we need simplifications. Simplifications are important because they make communication between doctors, between scientists, and between doctors and patients very easy. And so, one of the many simplifications that we use in clinical practice is chronic kidney disease. It’s a category; it’s a mental construct that we use for ease of communication.
The definition of chronic kidney disease is based on two things and two things only: one is GFR, that is a model of filtration—kidney filtration of poisons—and the presence of albumin or protein in general in the urine. So we basically what we basically do is we take a whole bunch of possible diseases that can affect the kidney and we characterize all of these different diseases based on just two things, okay: one is filtration, kidney filtration, and the other one is the presence of and the quantity of protein in the urine.
While albuminuria—protein in the urine—is very straightforward and can be readily measured from a urine sample, eGFR, that is the estimated filtration capacity of the kidney, is a complex concept that is an estimation of an estimation of an estimation. Theoretically, the filtration markers in general, and eGFR specifically, is a kind of a model of filtering capacity and therefore it very roughly characterizes the kidney. And so we use one specific substance called creatinine to see how fast the kidney is able to filter this substance and therefore we then say that the kidney is good if it filters creatinine fast, and the kidney is bad if it does this very slowly, okay.
Now, of course, the problem is that eGFR is a marker of filtering capacity regarding creatinine only, and we then generalize this as a model for filtration in general and of all functions of the kidney, okay. So because it’s such a crude estimation of kidney function in general, we should be very careful in the interpretation of eGFR. And yet, what we see in clinical practice is the exact opposite—that it is used for a substitution of proper diagnosis, and the diagnosis of chronic disease is usually made based on this number without understanding the background.
So let’s see how this eGFR is measured. Well, ideally, first of all, we need to understand that eGFR was originally used in physiology, in other words, in science. And whenever we take a scientific concept and try to use it in real life, then we generally are running into problems and very often these problems are not properly understood. So let’s try to understand. If we are we want to measure GFR scientifically, okay, in the context of physiology scientific measurement, then we have isotope-based studies. What we do is we ask the patient to stay with us for a whole day and then we take several measurements and we see at what speed a given substance, which is labeled by isotope and injected into a vein, is removed by the kidney, okay. So it’s gradually removed and how fast it is removed is a model of how good your kidneys are, okay. However, it is very time consuming, it requires intravenous administration of a special substance, and it requires multiple blood draws.
So you try to simplify the process by using creatinine from your own body. Creatinine is a substance that is normally produced by your muscles, so you don’t have to get injected by an isotope-labeled drug and no IVs, and it’s very much easier this way. And so we use creatinine to see how much and how fast you are able to remove creatinine by the kidneys. Okay, so far so good.
But this comes as a risk because the level of creatinine in the blood is not only a function of your kidneys, but there are other factors that influence what your creatinine level is. And here we go back to what I said before: so if we generalize this whole process and we want to say that creatinine in one person means exactly the same as creatinine in another person, then the individual uniqueness is lost. And individual uniqueness of patients can manifest itself in a myriad of ways. For instance, one patient may have a larger muscle mass, so that person will have more creatinine in the body independent of how good the kidneys are, right? If a patient has a muscle injury or is just generally sick and so the muscle is broken down because of an ongoing inflammation, then the creatinine is going to be higher independent of what the kidneys are doing because the kidneys are overwhelmed.
If a person eats more meat than another person, then the creatinine may be a function of what the person ate just before the blood draw, okay, with which we establish their creatinine level in the blood. So individual variety manifests itself in several ways so that there is a difference between an average of persons that were studied in a scientific study and the given patient in a given state, in a given condition, okay, and that needs to be very carefully interpreted.
Now, if you are a dialysis patient, then since unless your fistula is failing or your catheter is not working well, you are going to get approximately the same dialysis dose all the time. So what that means is that your creatinine is not a function of your kidneys or dialysis, but it is a function of these other things, okay. And so in dialysis patients, a higher creatinine often means that you are in a better shape; you’re eating more, you have more muscle, and you don’t have these acute problems.
Creatinine is filtered but it is also removed in other ways by the kidney, so it’s not really an ideal marker. And remember that creatinine is one substance that the kidney is working on—there are many other substances. So how well can creatinine be used to estimate one of the many functions of the kidneys? Well, so in addition to that, originally this was measured as a 24-hour urine collection, which is very cumbersome because you have to collect your urine precisely… it’s very imprecise. So this already a very imprecise model—this estimation of an estimation of an estimation—is further simplified by not doing this 24-hour urine collection and estimating what your 24-hour urine collection would show if you measured it, depending on certain risks.
Since people identified three risks as important, therefore it was examined how a serum or blood creatinine level predicts 24-hour collection results based on age, sex, and ethnicity. And note that ethnicity was here only in terms of African Americans and European Americans, okay. So we picked out three factors that influence how creatinine value in the blood determines a 24-hour collection result, and then we omitted the 24-hour measurement and we say that, “Oh, we are sure that the 24-hour will show this,” and so we don’t need to do it, okay.
So here’s the formula and notice here that eGFR is only a function of serum creatinine given a certain age, certain sex, and if you are African-American or not. What other factors were not examined? But there is more, okay. So like all mathematical formulas, this one is also general and it estimates things considering a limited number of factors, in this case three: age, sex, and ethnicity, okay. So how creatinine—the relationship of creatinine and 24-hour collection results correlates based on just these three factors.
But how precise is this going to be? Notice that these investigations were done in Americans—not in Japanese, not in Chinese, not in Hungarians—but in Americans. And these patients were usually devoid of kidney disease; they were healthy. Now you generalize something that was only determined in Americans and in patients who are relatively healthy to kidney disease patients. Now, how relevant is that? Well, you don’t know because the studies have not been done. So we are generalizing to all ethnic groups and all kinds of kidney diagnosis. This simple formula is very, very, very much estimated—idealized. It’s a model of a model of a model.
Now there are some obvious scenarios where this model is even less relevant than on average, okay. If you lack muscle, you have very little muscle, then how can you use a muscle degradation product to estimate kidney function? Well, not very well. If you exercise right before your blood draw, then your kidney is going to be overwhelmed by the amount of creatinine that comes out of an overworked muscle, no matter how your kidneys are, so it’s not going to be a good estimate of what your kidney can do. So if you have an obstruction to your urine, then of course your creatinine is really not going to be a good estimate. Well, what’s going to be a better estimate is how much urine you are able to make with the obstruction, okay. And if your urine is zero, then your eGFR is going to be zero no matter what your blood shows.
So here comes a wonderful idea. So we said that the problem was that you are considered an ideal average, right? So instead of taking a lot of blood draws from one person, you take one single blood draw from many persons, necessitating the idea that you are an ideal average. Okay, we reverse this and we are going to say that instead of comparing you to an ideal average, we are going to compare you to you, okay. So I’m gonna see what your creatinine is compared to what your creatinine was the previous month and see where it is going. We measure creatinine multiple times in the same person and see which direction it is taking.
Notice that since we compare one result and the next in a one person only, Beta Phi gr because age, sex, and ethnicity remains largely the same from one month to another, okay. And so we simply do not need this formula, this formula here which is based on age and ethnicity, because it remains the same. So it falls out of the equation, okay. So we just simply follow creatinine in the same person and see where it is going.
Notice that filtration does not model kidney disease that is a slowly progressive process because filtration is a function of many things, not just kidney disease. For instance, it depends on the actual state of the human body, so it is a snapshot, you see. So your blood pressure is going to influence it; your fluid status is going to influence it. So if you lose a lot of salt and water, or you have too much salt and water in your body and therefore your heart is not functioning well, then decreased filtration is not gonna tell you that you have kidney disease but only that your kidney is able to function less because of an acute heart condition for instance, or an acute diarrhea for instance, okay. So that’s not kidney disease; your kidney is perfectly fine, it’s just not given the conditions to function well at a given moment.
And of course diet and certain medications can influence this as well from day to day, you know, in a very temporary basis. We need to follow you up and see what happens with you, okay, because the more measurements we have, the less we are exposed to get a false reading. So we need multiple measurements. And yet, scientific studies usually have very few measurements of the same person in time. But in clinical practice, we are trying to do the exact opposite; in other words, we need to follow you up every month, every three months, every four months and then check which way your creatinine is changing, okay.
Now, that’s again so that we can sort out what is really happening with you at the particular moment. We need to be sure, okay, if you have an infection or not, if you have any kind of inflammation in your body. What is your actual heart condition? So if you have heart failure and you want to keep your kidney function stable and drink a lot of fluid, then you may go into fluid overload and that is going to mean not only a heart failure but kidney failure as well, okay. If you eat a lot of meat, I need to know about it because if your creatinine goes up, then maybe it’s more a function of your diet the previous day rather than what your kidney is able to do, okay.
Dairy products, amounts of salt and fluid in your body, okay, what medications you are taking… if you took a pain pill, I need to know about it because maybe what I see in the lab is the function of that medication, not what your kidney is able to do, and blood pressure changes. And notice that if you’re a diabetic or if you have some kind of blood vessel disease like hypertension or others, then all these issues are more of a problem because the kidney is able to compensate less. So that anything happens to you will change widely of what your snapshot blood value is going to be, okay.
Real kidney disease tends to progress slowly, okay. So if you have real kidney disease, then what I see is that your kidney function is less and less and less and less, and unless I start you on some kind of a treatment, it will just go to zero, okay. Now, acute injury to the kidney can cause a temporary decline or accelerate the original kidney disease, okay. Or if you have certain kind of risk factors—for example, you are diabetic—then acute injury may be more severe. Diabetes and blood vessel disease generally lower the compensatory capacity of the kidney, so the kidney becomes more vulnerable to injury making adaptation more difficult.
Also, certain medications which are otherwise very good for you because they prevent overworking in the kidney—over-filtration, hyperfiltration of the kidneys—okay, so that we need to really know all of these things and consider all of these things so we can interpret your actual blood result accordingly. Diagnosing kidney disease and predicting outcomes solely on the basis of a blood draw is like magic or weather forecast. You are subject to great errors if you don’t consider, if you don’t see the whole picture, okay. If you don’t consider risk factors and the actual state of the body which was prevalent when the blood was drawn, okay. So you need a well-experienced nephrologist with a great perspective so that this blood normal result can be properly interpreted.
Now what about prediction? Well, we are very confident that we can predict things, and so eGFR is this king and champion of predictions, okay. So if we have a few measurements, can we predict the results of future measurements? But sometimes yes, and the simplest is to draw a straight line. So if your parent was like this two months ago and like this one month ago, then there is a chance that it will be like this this month, okay. But the problem is that the actual situation is very rarely this simple, yet this is precisely what is done.
So the difference between mathematics and real life is that in mathematics a few fixed variables change and everything else remains constant, whereas in real life it’s the exact opposite: nothing remains constant because things change all the time. So straight lines rarely apply in real life. Diet changes, blood pressure changes, exercise, salt intake varies... blood work is a snapshot, so it counts what happens before an actual blood work. So if I don’t take into account all of these things and I make a prediction like this, then this can be very, very imprecise, okay, unless really everything is stable and you have a slowly progressive kidney disease.
But those are just assumptions, right? So what do scientists do? You see, if you’re not actually seeing patients but you are just kind of bogged down in theory, entangled in all kinds of great scientific concepts… instead of following one person, they measure a lot of people and then average things out. So they compensate a lot of measurements in one person following up that one person and understanding how certain things in that one person influence the progression of the result, they make that one person an ideal average of a sample population. So a lot of persons are investigated and then there is an idealized average of a lot of persons, and then an abstract idea is born from that average, idealized average. And that abstract idea is then forced upon you, right? So that’s what scientists do.
So everybody is regarded as an ideal average person; you, you, you, you, you—everybody—individual uniqueness is lost. A few fixed variables or known risk factors are considered in isolation. So that’s what holy science does to medical practice unless you really consider the individual differences in a given case. Time dynamics are ignored. How do we handle this? Well, we handle this by understanding that science provides us statistics based on imaginary homogeneity, and we understand that science gives us some ideas, but these ideas then need to be properly adopted in real practice when I am actually—this is a younger version of me—talking to an actual patient, okay.
So science does not say no, not at all, okay. Science needs to be carefully employed in your everyday practice with all the limitations understood. So people with a single low eGFR measurement can be falsely diagnosed with kidney disease, especially the elderly on certain blood pressure medications, okay. And there is all kinds of anxiety, etc., unnecessary diagnostic tests, anxiety, being labeled with the wrong diagnosis. But the reverse is also true, okay. If you just take one measurement, you may not be able to see that somebody has real kidney disease and it’s just slowly progressing, okay. The absolute minimum to circumvent this problem is to find an experienced nephrologist and tell him everything: in what state you are, what medications you are taking, what diets you are eating, okay. So instead of rough generalizations, you are going to get a specific interpretation of what a given blood draw result means for you, okay.
So here are some patterns. Kidney filtering capacity, which is not the same as kidney disease, it’s not the same as kidney function, but we just roughly generalize things and simplify things, okay. So kidney capacity tends to decrease with age, but at some point this decrease levels off if you are otherwise healthy, in other words, if you have a relatively healthy kidney, okay. So it’s not like I’m 75, I have an eGFR of, I don’t know, 20, and by the time I’m 95, I have an eGFR of zero. If that was the case, everybody would die of kidney disease, which is clearly not the case.
This decrease is accentuated with certain types of medications, and usually those kind of medications that are otherwise good for you because they prevent overworking in the kidney—over-filtration, hyperfiltration of the kidneys—okay. So your actual decreased filtration is good in this sense because then it would be more stable. Yet people who statically look at you might interpret this as you have chronic kidney disease and you are falsely labeled with kidney disease.
If kidney disease is really present, okay, filtering capacity will continue to decline, and especially if you have like a diabetic or you have blood vessel disease, this decline will be also more dependent on other issues because your kidney cannot compensate, okay. But unless your kidney function is measured frequently, you can have false projections. Why? Because if this is how your kidney function declines and you have two points, you see, like this and this one, and you insist like mathematicians do that it’s a linear decline, then you are going to predict this instead of that, you see.
The same thing if you have true kidney disease and your kidney function would go like this, but you start a treatment and you slow down and stabilize your kidney function and it’s going to go like this, okay. So everybody is different and everybody needs to be followed up carefully, and one number may mean something entirely different in one person compared to another and in one condition compared to another. Acute insults may worsen kidney disease like here, but these can be worse or they can even become permanent in diabetics or in certain patients with blood vessel or heart disease, okay. But there may be some recovery. And so in this situation you really need to follow up the kidney function, because if you have recovery, then you become too pessimistic not knowing that there is recovery if you don’t follow up the kidney function with multiple measurements, okay.
Now the problem is that unfortunately most non-nephrologists do not understand the delicacy of this, the multiple patterns that may be actually fitting your condition, okay. And so they generalize on the basis of scientific studies and they predict something wrong for you, okay. So those are the dangers of taking a concept—well, in this case a scientific concept—to medicine, okay, everyday medicine, everyday medical practice. And if you think about what happens in certain other situations where a general concept is not adapted to a local situation and it’s forced upon a local situation, then what happens then? But that’s not science and that’s not medicine, so we are not going to discuss that.
So what should we do with you, the patient? What should we do? Well, first of all, we make you understand that we should not mix up a snapshot of blood tests with true progressive disease, okay. And to decrease the influence of snapshots, we then interpret a given blood test result knowing your condition, okay, your actual condition of that particular time when the blood draw was taken, and we then follow you up and see where you are going, okay. And so we make predictions on the basis of following you up and talking with you multiple times. Follow up and let the nephrologist interpret and give advice, but the only way that the nephrologist can give advice is that you tell him or her what has happened to you: okay, if you took a new medication, or if your diet changed, or if you had a flu, or if you had diarrhea, okay. So we need to understand so that we don’t make false assumptions.
Elderly patients, especially on certain kind of medications like kidney-protective medications, blood pressure medications, etc., will have low filtering capacity without kidney disease, and the whole thing may be just very stable. And they are falsely labeled with kidney disease, creating a lot of anxiety and unnecessary tests and treatments. We realize that especially before blood work, diet, blood pressure, heart condition need to be stabilized; otherwise, your snapshot is not going to give you any predictive value at all. Vigorous exercise needs to be avoided, okay, in order to get a clearer picture.
And always remember that the conclusions based on some mathematical scientific concept not properly adopted in a certain situation given certain things that maybe change in you, okay, like for instance your daily diet, can result in false labels or, you know, on the other side of the spectrum, not noticing that you really have a kidney disease and the test just doesn’t show it, okay. In other words, what you need to do is to tell your nephrologist about your actual condition, take multiple measurements, and that nephrologist, based on all these data, will give you a prediction and change your therapy or medications or diet, okay, which is going to be good for you specifically, okay. And in the next presentation we’re going to talk about something very interesting: how we should change your medications on the basis of changes in the kidney filtration capacity. So can we use that to change your medications and how we do that? But until then, thank you for your attention and goodbye for now.
