Over the next 20 minutes, I'm going to get into some of the nitty gritty of invasive hemodynamic monitoring. I'll use a bunch of examples from our cardiac cath lab because it's easier to see what's going on there, but this all applies in the intensive care unit as well. Hello, this is Cliff Grayson again from the Rocky Mountain Regional VA Medical Center at the University of Colorado. I still have no disclosures. Formally, hemodynamics is the study of the forces involved in the circulation of the blood. But when we talk about hemodynamic monitoring, we're generally speaking both a little more specifically and a little more broadly. Practically, we're usually talking about measurement of venous and systemic blood pressure, of course, maybe blood volume, definitely blood flow, and importantly, metabolism and oxygen transport because, of course, that's what the blood is for. In the cath lab, we measure these things at a single moment in time. We obtained a single set of measurements or a snapshot, although occasionally we'll use some sort of provocative maneuver to gain more insight into what's going on. In the ICU, we're doing this with an eye toward deciding where we want to go. For example, if we want to implement some form of goal-directed therapy. Based on our measurements, we decide what's the best way of getting there. For example, using inotropes versus vasopressors or fluid boluses. And then we commonly make repeated measurements over time to find out if we're there yet. The first detailed measurements of blood pressure were probably made by this dour-looking guy, Stephen Hales. The woodblock print illustrates his account of his first attempt to measure blood pressure. A long glass tube was actually inserted into the carotid artery of a horse while the horse was restrained on the ground. Evidently, the Animal Care and Use Committee wasn't paying much attention. We've made a few gains since then. And of course, now there are myriad ways of getting at all this information. As you all know, there have been impressive advances in non-invasive techniques in recent years. There are those who assert that non-invasive techniques have made invasive hemodynamic monitoring obsolete. I beg to differ. Now I'll admit that one domain where non-invasive measurement predominates is measurement of systemic blood pressure. Commonly relied on various versions of a blood pressure cuff around an arm or leg. They're actually the gold standard in the ambulatory setting, mainly for practical reasons and because all the outcome studies were based on them. But of course, there are obvious limitations. For example, continuous measurement just isn't possible. And rapid repeated measurements are inadvisable and at the very least, uncomfortable. The pain of the cuff can transiently increase blood pressure. Patients with low cardiac output may not be easily measurable at all. Additionally, body habits can affect the accuracy, particularly of the systolic and diastolic estimates. While I still use a stethoscope and listen for carotid cuff sounds, these days oscillometric methods predominate. While these are pretty good for mean arterial pressure, systolic and diastolic pressures may or may not be accurate. And there's just something inherently distasteful about using a method where the algorithm is proprietary or at the very least, obscure. That said, for intermittent, infrequent measurements, they rule. Many of our echocardiography friends, and I'm an echocardiographer myself, will happily report estimated left atrial filling pressure based on echo. The best validated technique is based on the ratio of the early mitral inflow velocity to the mitral valvular annular velocity. We call it the EDE prime ratio. This graph comes from the foundational report that compared direct left heart catheterization and echocardiographic assessment obtained within a short time interval. This by the way, comes from the Mayo Group, one of the most respected echocardiography groups on the planet. The technique hasn't really changed much in the last 25 years. As you can see, there's an enormous amount of overlap in invasively measured EDP compared with the echo estimate. If you're doing research where you only care about the mean value from a group of patients, this is fine. When taking care of individual patients, not so much. And I'm sorry to say the state of the art is no more advanced now than it was when this was published in 2000. How about pulmonary pressures? Assessing PA pressure is critical to managing patients with respiratory failure and mixed valve disease. And as an echocardiographer, I'm asked to reassess PA pressure in patients with pulmonary hypertension routinely. This is a simultaneous right heart cath echo study from the Johns Hopkins Pulmonary Hypertension Group. While there's certainly good correlation on a group basis, once again, you can see that the estimates for individuals can be way, way off. But unfortunately, it's not the case that they're repeatable within subjects. Finally, what about every cardiology fellow's favorite middle of the night call? Could you come in and do an echo to tell me what the volume status is? The hospitalist is, of course, asking for an estimate of central venous pressure by evaluating the inferior vena cava. Unfortunately, while my colleagues in the echo lab confidently report this out, the correlation is weak at best, and by the way, completely unvalidated in patients on mechanical ventilation. When you put this all together, I'm pretty confident saying that typical non-invasive hemodynamic measurements are simply inadequate for developing an accurate picture of your critically ill patient's clinical status, and that invasive monitoring is still essential in complex cases, although even if you do have accurate information, whether it helps patient outcomes is a topic for another lecture. If we are going to use invasive monitoring, it's important that the data we obtain be reliable, and for that, you really need to understand some basic principles of measurement. We're going to spend some time on each one of these elements of the measurement chain. As I said before, I'll be showing you examples primarily from the cardiac cath lab. Even though many of these are not typical ICU measurements, they do illustrate the issues that you need to be aware of. We'll talk about problems that arise beginning with the catheters and tubes, the transducers, and the recording system. I'm going to touch on some of the pitfalls in measurement that can invalidate your conclusions, including calibration errors, problems with the measurement chain, and understanding what you should measure and not measuring the wrong thing. Most importantly, we're going to talk about what constitutes good and bad data. Every measurement needs a standard and a scale. For pressure measurements, we need a zero reference. The so-called phlebostatic axis is the reference point for all our pressure measurements. How many times have you walked up to your patient and found him or her sitting up at a 45-degree angle with the transducer clamped to an IV pole six feet away? If it's not a yardstick with a bubble level somewhere nearby, chances are good the calibration was done by positioning the height of the transducer by eye, and it's not necessarily been readjusted since the last time your patient changed position. This can easily result in 5 to 10 millimeters of mercury misestimates, which on the venous side of things is a lot. Blood pressure is pulsatile, and a key component of pulsatile systems is the modification of the pressure signal that occurs before the pressure is even seen by a transducer. You may recall the term damping. This refers to a blunting or smoothing of a signal. If we take the green signal here as a true measurement, the blue is what you get when there's too much damping or smoothing, and the red signal is what happens when resonance or ringing contaminate and amplify the pulsatile components of the signal. The fluid columns we commonly use to measure pressure contribute to these effects, but these effects occur within the circulation itself. You can see that pulsatility of the systemic circulation varies dramatically as you move from the ventricle on the left out to the peripheral circulation on the right, as illustrated in this chart. There's actually amplification of the pressure signal by the arterial system in smaller arteries before resistance vessels begin to damp out the pulse. This is a simultaneous measurement of aortic pressure obtained from a pigtail catheter sitting in the ascending aorta compared with the pressure obtained from the femoral sheath. While the overall measurements here don't differ that dramatically, you can see the shape of the arterial waveform is different, with a dichrotic notch clearly visible in the aortic but not in the femoral measurements. The signal has been slightly smoothed out or damped. Here are some real-world samples I obtained in our catheterization lab. This is a high-quality arterial waveform obtained with a decent caliber pigtail catheter sitting in the ascending aorta. This is what it looks like after the saline in the catheter has been flushed with radiographic contrast agent. The mean pressures are similar, but the systolic pressure is lower and the diastolic pressure is slightly higher. In the arterial system, we may not care that much about the systolic and diastolic pressure For example, this is a left ventricular waveform measured with a large-caliber pigtail catheter. Here's the same ventricle after radiographic contrast agent was introduced into the catheter and connecting tubing. It doesn't look much like a ventricular waveform anymore, does it? You can see similar effects if blood clots or air get into the catheter or tubing of your measured ventricle. Here's the same ventricle after radiographic contrast agent was introduced into the catheter and connecting tubing. Here's the same ventricular waveform measured with a large-caliber pigtail catheter or tubing of your measurement system. The solution to this is carefully flushing your system with saline, but you need to recognize the problem first. Overdamping is a more common problem than underdamping, but underdamping can be a problem too. This shows simultaneous left and right ventricular pressure tracings from a patient with constrictive pericarditis. The waveform on the left is a ventricular waveform I obtained from an experimental animal, a pig, using a micromanometer solid-state transducer catheter, where damping and resonance effects are minimal. The tracing on the right was obtained at the same time using a short length of large-bore pressure tubing. You can see the overshoot and undershoot in the waveform from resonance or amplification effects. Artifacts like these in RB pressure waveforms can be and frequently are misinterpreted as the square root sign of pericardial constriction, but are entirely due to underdamping or resonance. This problem can often be corrected by injecting an air bubble into the monitoring line. If you're uncomfortable with having an air bubble in the arterial system, you can use a stopcock to isolate it to a side port. This shows how the arterial tracing changes as increasingly large amounts of air are added to the side port of the stopcock. There are formal methods for evaluating the damping and resonance characteristics of a monitoring system, usually using a brief high-pressure saline flush. But you can usually get a pretty good idea of what's going on simply by recognizing what a normal tracing should look like. Let's move on from problems in the measuring system itself to problems with knowing what we're measuring. We commonly use right heart catheterization to assess patients with pulmonary hypertension. Here's a patient I took care of with acute right heart failure suspected to be from thromboembolic disease. You can see the flattened left ventricular septum in panel A, the large RV and RA in panel B and D, and the high tricuspid regurgitation jet velocity in panel C, indicating severe pulmonary hypertension. Based on the way the LV is being squeezed, you might think this patient has a pulmonary vascular abnormality rather than left heart disease. Here's the initial set of right heart cath measurements. You can see the very high RV pressure in panel A and the high PA pressure in panel B, with moderately elevated RA pressure in panel C, indicating right heart failure. These are all high-quality pressure tracings. These are all high-quality pressure tracings. But unexpectedly, the wedge pressure tracing in panel D shows a high pulmonary capillary wedge pressure. This would have led us to conclude that left-sided heart disease was driving his pulmonary hypertension. But the echo findings seem discordant. Why would the LV septum have been so flat if left ventricular and diastolic pressure were high? We can get some clues from thinking about timing. Recall that it takes time for pressure waveforms to propagate through the vascular system. This shows the time delay between measurement in the aortic route, the black tracing, and in the femoral artery, the red tracing. There's a good 50 to 100 milliseconds there. A similar time delay occurs in the pressure measurement from the left atrium obtained with the PA catheter. We can exploit this to determine if our pressure measurement is plausible. A more detailed examination of our wedge tracing provides an important clue. Notice that the peak of the presumed wedge tracing, marked as the V wave by the computer, occurs just after the T wave of the ECG tracing. But the peak of the pulmonary artery tracing does, too. This is not what you'd expect. The peaks of these waveforms should not occur simultaneously. This is actually more consistent with an incompletely wedged catheter and a damped PA tracing. Recognizing the problem, we repositioned the PA catheter, and we're able to obtain a true wedge pressure, showing the expected low left-sided filling pressure and confirming pulmonary arterial hypertension rather than pulmonary venous hypertension. When you get conflicting data, it's your job to try to understand the reason for the conflict. Here's an example of a patient being evaluated for suspected severe mitral regurgitation. The echocardiographic finders were equivocal. In this case, the invasive measurements were very helpful. He has moderate to severe pulmonary hypertension and a large V wave in the pulmonary artery wedge tracing, attributable to his severe MR with a relatively small, noncompliant left atrium. We used two techniques to assure ourselves that the tracing was actually a wedge tracing and not just a damped PA tracing. First, you can see a distinct delay in the peak of the wedge tracing V wave compared with the peak of the PA pressure tracing. In addition, we obtained a blood sample from the distal tip of the PA catheter and measured an oxygen saturation that was nearly equivalent to his systemic oxygen saturation, confirming that we were sampling only from the pulmonary venous side of the system. So that's pressure. What about flow? We have two main techniques, indicator dilution and FIC. Most of the devices sold by various manufacturers are just variations on these, and I won't go into them in detail. Indicator dilution is just a fancy name for mixing. The basic principle is that you take a known quantity of an indicator, mix it with an unknown volume, then measure the new concentration. Indesign and green was used traditionally, but the standard now in the cath lab and the ICU is typically a bolus of room temperature saline, and concentration is nothing more than measurement of temperature. We inject a bolus of room temperature saline, then measure the change in temperature of the downstream blood over a short period of time. Continuous cardiac output devices sometimes use a little heating coil to provide a continuous bolus, as it were, of warmed blood, and the dilution of heat is used to estimate the flow. It works so long as the heater isn't in contact with the vessel and the mixing is thorough. While the specific algorithms used are often proprietary, the general principle is the same among devices. High blood flow causes a rapid fall and rise with a small temperature change, while lower flow causes a slower fall and rise with a larger temperature change. Understanding the technique helps troubleshoot when you're getting crazy results. If the concentration or the quantity of the indicator are wrong, you'll get wrong results. This can happen because the temperature is wrong or the syringe was misfilled. Other common problems are shunts, leading to lost indicator, using the wrong port for injection, more common than you might think, and using too slow an injection and not getting enough signal. The other chief method of flow measurement is the Fick principle. This is based on the simple idea that the oxygen delivered by the arterial circulation has to equal the oxygen returned by the venous circulation plus the amount extracted by the body and presumably used for metabolism. Once again, understanding the principle can help you troubleshoot. Fick measurement is really easy. It's only dependent on three factors, oxygen consumption, oxygen content of blood in an arterial sample, and oxygen content of blood in a venous sample. The largest sources of error come from sampling oxygen content in the wrong place, for example, using peripheral rather than central veins, and mis-estimating oxygen consumption, which we don't often measure directly in the ICU. How many of you know what a Douglas bag is? Oxygen consumption is often much lower than commonly used equations estimate and can change markedly with body composition and disease state. By the way, mixed venous oxygen saturation, when obtained in an arterial sample, can be obtained from a central vein and preferably from the pulmonary artery, where everything is pretty well mixed, is a pretty decent way of estimating circulatory adequacy. If your mixed venous PO2 sat is much below 55% to 60%, you're probably in trouble. There are a few questions where we don't actually need invasive measurement. For example, I'm commonly asked to rule out intercardiac shunt in patients with hypoxemia. Don't forget the 100% oxygen test. If oxygen sat normalizes or is near normal with 100% inspired oxygen, your patient's hypoxemia is not due to a PFO or other intercardiac shunt. And actually, we don't usually need a central venous catheter to determine central venous pressure. JVD assessed visually or with the aid of an ultrasound device actually works pretty well and is better than echocardiographic assessment of IVC dynamics. You may not need a central venous catheter to estimate central venous pressure, but I've lost count of the number of times I've been asked to evaluate IVC dynamics for volume status, then walked into the patient's room only to find a central line that's never been transduced. If you have it, use it. It's the gold standard. But you do need to review the tracings properly. It's not easy to make out the different components of a central venous pressure tracing here, in part because the scale has been chosen badly. Simply setting the scale to something more suitable brings out the features of interest. Also, it's important to make measurements at the right time. For intravascular pressure measurement, we typically measure at end expiration when intrathoracic pressure is neutral compared with surrounding atmospheric pressure. For example, this patient's wedge pressure is 6 to 8 millimeters of mercury, not 1, as reported by the computer. The measurements need to be made at end expiration. Compared with pressure and flow, which could be accurately determined invasively, volume status is a much less developed concept. You'll hear about various ways of assessing fluid responsiveness. But don't forget that invasive measurements can be helpful in measuring response to interventions, such as volume infusion and use of vasoactive agents. Not uncommonly, we get unexpected or discordant results. It's usually due to one of three factors. First, there was a technical problem with the measurement. But the more you know about the technical details of your measurement technique, the better position you'll be in to understand when this was the problem. Secondly, your hypothesis or your diagnosis could have been wrong. That's why we test. Don't dismiss an unexpected result if there's a plausible explanation for it. Finally, be sure you understand the physiologic basis of the measurements you're making, and that you aren't just missing the point. Thanks.

invasive hemodynamic monitoring

blood pressure

blood flow

oxygen transport

calibration errors

volume status assessment