SCCM Resource Library-Continuous Versus Noncontinuous Glucose Monitoring

Video Transcription

Today, we are going to be talking about glucose monitoring, continuous versus non-continuous. My name is Judy Jacobi.  I'm a critical care pharmacist and I'll be your moderator for today's program. I have been involved in the glycemic control guidelines from the Society of Critical Care  Medicine and currently am co-chair on the revised guidelines on glycemic control for critically ill children and adults.  When we first published in 2012, we started out hoping we could talk about glycemic control, but the literature was not adequate to really help us narrow in on that topic, and so we  focused extensively on insulin infusion therapy. But we're able to make some comments about monitoring.  In the grade lexicon, these are all suggestions for most patients on an insulin infusion, but it's a suggestion because of relatively low quality of evidence.  We suggested that blood glucose be monitored every one to two hours, that point of care meters were acceptable, although not optimal for routine testing, that certainly arterial  and venous whole blood could be sampled instead of finger stick capillary sites, and that was especially important for patients in shock with poor perfusion, significant peripheral  edema, or prolonged insulin infusion. Further, we attempted to evaluate the literature on continuous glucose sensor utilization and  found inadequate information to make a suggestion for or against those sensors. In those eight years since that first guideline was published, certainly we find that perhaps  insulin therapy goals have changed. We certainly have been challenged to manage patients and glycemic control in a COVID era  with minimal PPE and need to avoid frequent patient contact, and hopefully we'll find with the help of our speakers today that technology has advanced as we hope.  We have some terrific speakers today. We have Michael Agus, a physician from Boston Children's Hospital who is an expert and researcher  in pediatric endocrine disorders of critically ill children and was involved in the earlier as well as the current guideline update.  Ruth Leavitt from the Children's Hospital of Philadelphia talking about the nursing burden of bedside monitoring and titration and certainly pertinent in a COVID era, and  Rochelle Firestone talking about tricks and traps, other factors that affect glucose monitoring.  That will hopefully accomplish our objectives to talk about current technologies for glycemic  monitoring, talk more in depth about the workload effects and how it can manage, how we manage  and control our glucoses, and certainly medications and other factors that will influence glucose management and control.  I'd now like to turn the podium over to Michael Agus to talk about technologies in glycemic control. Hi. My name is Dr. Michael Agus.  I am a pediatric endocrinologist and critical care doctor at Boston Children's Hospital.  I appreciate the opportunity today to talk a little bit about technology in glucose monitoring.  My disclosures are that in the past, we've had some discounts on study devices provided to us by the various companies in our controlled trials, including Dexcom, Medtronic, Nova,  and Edwards, as well as grants from the NIH and our own Boston Children's Hospital. And finally, I've recently been a consultant at Roche Diagnostics.  We will cover some off-label topics today, including hospital-based subcutaneous continuous glucose monitoring, CGM and non-diabetics in young children, and in fact, IV insulin  in the sense that only one of the companies has actually approved IV, gotten FDA approval for IV insulin in our inpatients.  I'll start off with my own bias, which is concerns about hypoglycemia. As a pediatric doctor, I worry about the cognitive outcomes of our patients in the ICU.  For me, the primary question is how low for how long? Is glucose acting similar to oxygen in that low glucose produces injury?  We just need to know as best we can what the blood sugar is in order to avoid that hypoglycemia.  There's plenty of research on hyperglycemia in terms of glucose control, but the predominant injury that we see seems to be more related to hypo.  So with that bias understood, I'm going to jump into the different approaches to glucose monitoring.  Now, our first decision, as always, do we look at whole blood versus serum versus interstitial fluid?  I'm not going to go deep into the data given our short time today, but to tell you that there has been an enormous body of literature looking at these differences.  They are small and significant from a lab point of view, but I think there is general clinical consensus that they are not significant from a clinical point of view, and that we  use whole blood devices, compare them to serum, even though there may be between a 5% to 10%  difference in interstitial fluid. We'll talk about the limitations in a little bit. The gold standard in most glucose monitoring studies has always been the YSI device, and  that generally is considered to outperform only the blood gas analyzers, which can be at the bedside, the unit, or the central lab, or central lab analyzers, which generally  measure serum levels of glucose. In the hospital, the Novimeter is the only one approved for ICU use using a glucose oxidase  enzyme, and Roche approved for use in the rest of the hospital using glucose dehydrogenase.  We know that arterial samples with the appropriate waste, meaning a non-diluted sample, is going  to be our top choice, but venous free-flowing and capillary free-flowing are also very close to arterial concentrations.  The Novimeter, which I mentioned, is the only bedside glucose meter I'm going to show you today because it's the only one that is approved for use in critically ill patients by the  FDA. The company has also gone looking for interference, meaning medications that are present in ICU  patients that might interfere with glucose measurement, and across 53 studies so far have not detected any of those interferences.  Now the question of arterial versus venous has been on our minds for many years. 1985 years ago was the first study that I could find in the Indian Medical Gazette,  which determined that there was a 3 milligram per deciliter difference between arterial and venous. I'm not going to go further into that other than to say what I said earlier.  These differences are not considered to be clinically significant at this point.  Jumping into continuous glucose monitoring, just to show you some performance in case you haven't seen these sorts of data in ICU patients, and this is the world that I live  in, in children, the red dots are blood gas measurements, and then the lines are the CGM, subcutaneous or interstitial fluid measurements.  And as you see, just pretty impressive performance. And so we'll use that as our baseline to say that CGM really may have some value in our patients.  Now if we look at the concept of continuous glucose monitoring, it's not only interstitial fluid.  In fact, there's a whole host of approaches that may be invasive, meaning placing needles or catheters using microdialysis, micropores, microneedles, subcutaneous amperometric  electrodes, that's our interstitial fluid CGM devices, and then intravenous implantable devices, which were a hot topic several years ago.  Non-invasively, there are a lot of companies looking at a lot of different approaches. I won't go through those other than to say these have not yet come to market.  Companies have been working for several years and spending a lot of resources to get there, and it'd be wonderful if they did, but to date, they have not.  So the two approaches that seemed most promising are the subcutaneous electrodes and microdialysis,  and predominantly the subcutaneous electrodes, of course.  The benefits of using interstitial fluid versus blood in terms of leaving a sensor in for  an extended period of time are reduced biofouling. You don't have fibrin deposition. You don't have clots on the sensor if you're not in the vascular space.  You do have the issue of time lag, but that time lag has really been shown in many studies  to be very short and, again, probably not of significant clinical significance, especially when  rapid changes induced mostly by feeding are less common in our critically ill patients.  In terms of what we are allowed to use in the hospital, there are two FDA-approved systems currently, one called the GlucoScout and one, the OptiScanner, both of which use  venous drawing systems I'll show you in a second.  In Europe, the OptiScanner is also approved, and there were four other devices that have  been used over the last several years, but unfortunately, none of them is currently available.  OptiScanner is a device that essentially uses a blood draw that is automatic by the device at the bedside through a peripheral IV.  It draws about three mLs of blood, and then it's spun down to about 120 microliters  of serum, which is then measured by the device and reported.  Similar approach is with the GlucoScout, before it used to be called the VIA device,  which again draws blood out of the IV system, measures it, and then returns it to the patient.  The problem in all these devices is these mechanical malfunctions, and these are examples where the sample delivery was incomplete to the home device that's attached to the patient,  but you can imagine even the complexity we have in getting a human to draw these  samples perfectly in terms of clotting, in terms of blowing the IV, entraining air, faulty connections, that sort of thing. Not to mention these devices need to be  attached to the patient. A simpler, cleaner device may be a subcutaneous,  very small portable device like the ones we use in ambulatory diabetics.  And you'll see over the years, starting in 1999, so now 22 years ago, we had our first approval  of a subcutaneous device. They were initially in adults, and now they're increasingly cleared for  young children down to two years of age. And the most recent devices in each of these companies requires two and often no calibration, a very brief warm-up, and can stay in for many  days. All those details are slightly different for each model. What's amazing is that there's  four on the market at present. The Guardian by Medtronic, the Dexcom G6, the Abbott Libra 2,  as well as the Libra 14 day, and Sensonics ever since.  The first three are the ones that are also used on ambulatory diabetics. And in April of 2020,  the FDA notified these companies that it was acceptable for them to be used in the hospital.  Not critical care areas named per se, but people have been using them. And so there's a real  opportunity to bring this forward. I showed this microanalysis device. It's unfortunately no longer available, so I won't go into great detail, but that uses slightly different  technology, obviously, than these electrodes. The electrode is an enzyme-coated electrode, which is inserted into the subcutaneous tissue. And the chemical reaction, it produces  electrical current, which is sensed and relayed back to the monitor, which estimates the blood  glucose. Now, there's multiple electrodes. And on the electrode, there are these various membranes  which block permeability to some of the medications we use in the ICTU. This is the only subcutaneous device that was originally built for inpatients. Again, no longer available.  But interesting in that it addressed the issues of screen, the sound, making a good alarm for  bedside nurses, and looked at the drug interference as a very important issue in the ICU.  The other devices you've seen, the Medtronic, very easy to insert, the Dexcom, they all talk wirelessly now to Bluetooth devices. Young children are able to insert  on their own, and so certainly bedside staff should be able to insert these.  The Abbott approach used to be that you needed to hover over the device in order to get a reading,  but in their most recent version, you see trends with continuous, or intermittent, but frequent communication between the device and the sensor. And in fact,  there's platforms where you can measure the blood sugars of multiple patients at the same time.  In the hospital, we know we can use it for hypoglycemia, as I said,  as a pediatrician, that's really my primary concern. But we've done plenty of work on hyperglycemia, and in the adult environment, hyperglycemia is a more dominant concern.  In diabetes, DKA, but also just run-of-the-mill critical illness, hyperglycemia. And in all age groups, these devices can provide advice on both sides of the spectrum.  We did our own prospective trial in almost 1,000 babies after cardiac surgery, and got very good control using a Medtronic subcutaneous sensor. We had an algorithm at  the bedside that helped us guide how much insulin to give to get the desired glucose range.  We also use, which is very interesting, an important consideration, if you're drawing blood glucose levels, is some sort of closed device. This one is called the VAMP,  but some sort of closed device to get rid of waste responsibly, and in this case,  it stays in this area of the system, so it can be returned to the patient. But the idea is you  have a very clear approach of avoiding dilution in your blood measurements, and it improved the  performance of the system quite substantially from an MARD of 10 down to 6. The performance of the subcutaneous sensor was quite good, although our MARD, our mean absolute relative  difference, which is kind of the overall score of glucose sensor performance,  was kind of medium in the 16 range. In our second trial, where we did about 700 patients, non-neonates, going prospectively across three different sensors, we had really significant  improvement down to 12% in our MARD. And although we had false positives and false negatives,  the data suggests that even in these false positives and negatives, they were quite useful from a clinical point of view in determining directionality and avoiding emergency events.  Now, we were able to detect almost half of impending hypoglycemia events, even with a sensor  that was performing in not the best fashion at 17.6% with an MARD. We were able to reduce our severe hypoglycemia rate down to below the previous baseline, even when we're giving  insulin and targeting low values. And so these CGMs really can enhance avoidance of hypo and  control of hyper. We did not see effects in terms of edematous patients. Medications, some are quite  important. Acetaminophen is being screened out in the most current versions of CGMs. Perfusion seemed to be an issue in very low cardiac output states, like in ECMO patients  on very high doses of pressors. But in the great majority of patients, perfusion was not an issue.  Temperature does seem to be an issue, but only in severe hypothermia, as in in the operating room.  In very small babies, we've even used these. This is a group in Italy that used the CGM in preterm babies. And another group used these to predict complex events in the NICU.  And so we have to think about the warm-up, the point accuracy, of course, the value, what are  we using this value for, and even issues like the display and the sound in our busy ICU environments.  We know, as I said, that there are some issues with low cardiac output, vasoconstriction, body wall edema, less of an issue, temperature, and then, of course,  the effect of drugs in giving us false appearing elevations. But what's important to note is that  even when our MARD was in the higher ranges, we were still able to achieve really impressive reduction in hypoglycemia and control of hyperglycemia. Now we have performance  values that are even better. And if we look at the most recent sensors, they've really broken  this 10 percent MARD barrier. And three of the devices are below 10 percent, 8.79 and 9.2 percent,  really outstanding performance. And so we've gotten to the point where we really have a device  that if FDA allows us, which they currently are, can really allow us a great improvement  in the care of our patients. And so I think we're at the point where our CGM manufacturers need to  really gather the data in the ICU and allow us official guidance and approval to use it, and the  clinicians need to consider embracing the technology and bring it in-house. We've learned  through this process that the perfect MARD may not be achievable, but even with MARDs that are less than perfect, we can achieve a really significant enhancement in the care of our  patients. An enormous number of people have contributed to collecting all these data.  I appreciate the work of all of them over many, many years. Thank you very much for the opportunity.  Good afternoon. My name is Ruth LeBay. I'm a nurse scientist at the Children's Hospital of  Philadelphia. I also teach in the Clinical Nurse Specialist Program at the University of Pennsylvania  School of Nursing. And my presentation, my contribution to this seminar is to talk about the nursing burden of bedside monitoring and titration. I have nothing to disclose, and  my learning objective for my part of the session is to review factors impacting nursing burden related to implementing glycemic control protocols.  So, protocolized care is a very important part of ICU management of patients, and ICU nurses  manage between 5 and 29 protocols per patient each day. So, protocolized care really impacts  the care that patients receive, and it really impacts the work of the nurses because nurses  play a key role in successfully implementing these protocols, which heavily impacts patient outcomes.  Although intended to streamline nurses' work, protocols can add to nursing workload,  particularly cognitive burden, especially when they are attempting to balance and juggle multiple  protocols. Glycemic control protocols are very widespread, and studies show that nurses perceive  increased workload related to glycemic control protocols, and that generally is most specifically related to frequent glucose monitoring and infusion rate changes that are required.  As one nurse in a study that I'll be discussing stated, I understand the potential benefit of glycemic control, but I felt like one nurse's assignment today could have entailed just  checking the blood glucoses and titrating the insulin drip. So, you can see that that is really  a heavy focus of the nurse's work related to glycemic control protocols. So, why is it useful  to understand the nursing workload burden related to glycemic control protocols? First of all,  the person who has the primary responsibility for implementing these protocols is the nurse.  So, it's important to understand the protocol's end-user perceptions, as how the protocol  is perceived will impact how it's implemented. Also, it helps to identify protocol pain points  or areas where there could really be improvement, which would both decrease nursing workload burden  and improve patient outcomes, and then it also allows us to design effective interventions that  decrease that burden and also, therefore, increase adherence and improve patient outcomes.  So, there are a few ways to measure workload burden, but one that has been used extensively  in research looking at nursing, physician, and pharmacist workload is the subjective workload  assessment technique tool, which measures three dimensions of workload that affect performance.  Three are cognitive burden, which is the amount of mental effort or concentration required  for implementing the protocol, the time burden, which is the presence or absence of spare time,  the amount of interruptions, and the task overlap that affect work performance, and then the  psychological stress burden, which is the presence of confusion, risk, frustration, and anxiety  associated with the work that affect work performance, and each of these is a key part of any workload. A second tool that has been used fairly extensively in medicine, nursing,  and pharmacy as well is the NASA task load index or the NASA TLX. This is a slightly different  tool. It's a six-dimension tool used to obtain an overall perception of workload related specifically  to high-stress tasks, of which I would argue that implementing a tight glycemic control protocol  in an ICU is definitely a high-stress task, and the six elements are the cognitive demand, the physical demand, the time pressure, the performance, or how well the individual is  able to accomplish their tasks, the effort in getting that high level of performance, and then the element of frustration. Cognitive burden is a really high part of the workload  because there's a lot of cognitive work that goes with any aspect of providing care in an ICU,  and that's true whether it's a nurse, a pharmacist, a physician. It results from the complexity of  new information to be learned and other activities that must be completed unrelated to a task you're  trying to complete, as well as the cognitive processes required to complete the task.  The burden increases when there's a change in the focus of attention, and this is particularly  related to interruptions or distractions, of which we know in the ICU environment, interruptions and distractions are high.  Individuals have a finite working memory capacity, and that means that the amount of information that anybody can take and process and remember is limited.  And what is thought to be the case from human factors work is that individuals can hold between five and nine tasks in their mind at any one time.  Thinking about a patient's care in the ICU who's on a tight glycemic control protocol, you know that five to nine tasks does not encompass the range of activity that's going on.  And with these multiple competing priorities, nurses must choose where to focus their attention and what tasks and information they will remember.  So I'm going to talk about nurses' perception of workload burden in pediatric critical care, which describes one aim of the Heart and Lung Failure Pediatric Insulin Titration Trial,  fondly known as APPLI. And the nursing aim was to investigate nursing workload related to implementing a tight glycemic control protocol.  It involved a bedside computerized algorithm incorporating continuous subcutaneous glucose and intermittent bedside glucose monitoring and provided ongoing protocol instructions.  The computerized decision support system theoretically decreased cognitive burden by completing calculations,  recommending adjustments of the insulin infusion, and providing a countdown to the next blood glucose measurement.  This decision support tool was incorporated into the nurse's workflow by being placed at the bedside so it was immediately available when a management decision had to be made,  and it required minimal data input and provided specific recommendations for next steps. There were 35 enrolling sites, which trained 3,278 nurses on the protocol.  Of that group, we got 1,476 to complete baseline surveys for a 45% response rate. And for 505 of the 693 randomized patients, we had completed intervention surveys, so  we had a 50% of the patients we have results for. We used two tools to measure their workload perception related to the tight glycemic or to the glycemic control protocol.  And nurses cared for patients who were randomized to a tight, very tight, or low target and a high target group.  And the comparison of nurse perception of workload by the workload dimensions of the SWOT between the lower target and the higher target group show that in terms of all three  elements, nurses who cared for patients in the tighter, the lower target group, perceived increased workload.  Similarly, when we measured those same workload burden perceptions using the NASA TLX, we found the same results.  Those in the lower target group, nurses cared for patients in the lower target group, perceived higher workload.  But what's really interesting is that the target range impact their perceived workload.  So the factors contributing to workload differed by the perceived patient risk and the vigilance required by the protocol.  So nurses caring for a patient in the lower target group reported higher cognitive, time, and psychological stress burden and had a higher overall paired SWOT score, indicating  that they overall saw a higher perceived workload when they felt their patient was thicker, more critical, and more at risk.  The results for the NASA TLX also showed higher perceived workload for the lower target group with all elements significantly higher.  So these two tools, which aim to measure the same thing, showed consistent results. When the glycemic control target was low, then the nurses perceived higher workload  and higher concern for their patient. So although nursing activities were the same in both groups, there were significant differences  in terms of which protocol elements contributed most to workload. So the nurses implementing the lower target identified more difficulty managing a patient  on the protocol, keeping the blood glucose levels within range. They had more concern about hypoglycemia, and it required more concentration, cognitive effort, and attention.  And the reason for that was that they were more concerned about the status of their patient, and also focused more on patient-specific issues.  They also perceived repetitive blood draws, adjusting the insulin infusion as their highest  contributors to workload, and the higher target group reported managing the glucose sensor using the computer as the two elements that had the highest impact on the workload.  Because they had less concern that their patient was more at risk, they focused more on the technology and how that impacted their work.  And it's interesting to note that the perception of workload did not change with nurse age, years of picky nursing experience, number of times the nurse had cared for a patient  on the protocol, or comfort level or attitude toward technology. So this was very consistent across the range of nurses that cared for this group of patients.  The findings were very consistent. Nurses reported the use of technology increased workload. They frequently reported difficulties with the continuous glucose monitoring sensor,  which is similar to other reports in the literature of how nurses perceived their workload. When there was increased technology, there was increased workload, whether it was related  to the patient specifically or managing the technology. And nurses perceived an increased workload related to protocol implementation, which  is consistent with findings of other studies as well. In this particular study, compliance was very high with the protocol, which differs from  previous studies, which found decreased compliance when patients were critically ill, but nurses  still had a very high concern for their patient based on how at risk they felt their patient was.  We have direct feedback from nurses, which supports the concerns that nurses report related to protocol implementation when technology is involved.  And I put some of these here for you to review. It's worth noting that things that made it harder to use the technology or time limitations  when nurses already felt stressed in terms of their time management really made protocol implementation much more challenging for them.  And that when patients have rapidly changing statuses, but the protocol is very time driven, this also represents a real challenge for nurses.  Some of what's important to know about implementing a glycemic control protocol, regardless of  whether you're using continuous monitoring or intermittent monitoring, whether it's paper driven or whether it's technology driven, is that you have to know how the technology  that you're introducing interfaces with other technology that the nurse has to work with. And constant recalibration and failure of technology really increases nurses' frustration.  One of the things that we very consistently saw with feedback from nurses in this study was that calibration of the machine and their lack of confidence in the accuracy of the  continuous monitoring sensor really, really was a factor. Other things that nurses recommended were what they perceived could make their work  easier, such as sending information to the laptop automatically rather than nurses having to enter data, which also could introduce error.  And again, feelings of being unable to accommodate the time constraints. But the nurse perceived that when she went from the continuous glucose sensor and entering  data to intermittent checks, that the time burden and the stress burden did decrease, which makes sense. This was a very, I think, a very telling comment from one of the nurses.  And I think it really encapsulates the benefits and the drawbacks of a technology driven nurse protocol for glycemic control.  Because as the nurse said, I love the idea of a nurse driven protocol for titrating infusions for a target goal.  It saves time, having to continually contact the physician every hour for orders to make changes. I really love that part.  However, RNs at my hospital are not allowed to run the bedside glucometer, so I would have to wait for someone else to run my values, which meant I frequently got behind on tasks.  And that I had a patient who got randomized to the tight glucose range while receiving steroids, which we all know impacts glucose levels, which meant that I was really chasing  changes all shift. The sensor never correlated very closely with the actual glucose, so I was frequently recalibrating.  The basic idea is good, but the implementation often feels stressful. So when implementing a glucose control protocol, you need to look at where are the pain points,  what will improve the protocol to ensure that the goal of the protocol is met, because it  realistically fits in with the work the nurse is doing and all of the other competing priorities  that the nurse has to manage in ensuring that she or he provides excellent patient care for their patient, which really is the focus of the nurse's work and really came through  as a theme in the study that we did related to tight glycemic control. Thank you. Good afternoon, all.  My name is Rochelle Firestone, and today I'll be discussing medications and other factors that affect your glucose when using new diabetes technology. I have nothing to disclose.  I'll say this presentation today will be through my lens as a practicing adult critical care pharmacist.  The learning objective given for this talk was very broad, and there's a lot to unpack here.  I will be highlighting a few of what I think are the more important factors that influence  glycemic control when using diabetes technology and what to do when you encounter these factors in your ICU.  So ideas about glycemic control in the ICU, as we all know, it's a bit of a pendulum, and there's still a lot of controversy that exists in this realm of the literature.  It is indisputable, however, we do know that dysglycemia is associated with poorer patient outcomes. But there remains a lot of questions.  For example, should we treat everyone that has hyperglycemia? If so, what blood glucose target is best? And how exactly do we do it? Do we protocolize it?  Should it be patient or population-specific? And how does continuous glucose monitoring technology fit into this picture?  We do have guideline recommendations specific to critically ill patients to help guide us. And these are from 2012.  And they suggest that IV insulin infusion therapy is preferred for most of our patients. However, we all know that this therapy comes at a cost.  Every one to two hour point of care checks are really burdensome for both the patient and the staff.  And it's really just a snapshot in time, like a single point of data in an otherwise very complex system. Glycemic excursions in the ICU are really multifactorial.  And they include things like parenteral and enteral feeds, medications, comorbidities, severity of illness, patient-specific metabolism. And this is a very broad topic.  We won't have time to discuss each of these in depth. So I will be touching upon a few, beginning with one of the main limitations of the device technologies that we use.  And that is medication interference. So jumping off with continuous glucose monitoring systems, or CGM, these are very popular in the outpatient setting.  They report blood glucose as well as the rate and trajectory of change, which is all information  that sounds like it would be very useful to have in an ICU setting to help inform insulin dosing and to prevent hypoglycemia.  But currently, use in hospitalized patients remains investigational. And there are some barriers that exist to their adoption.  Most specifically, technical limitations, like the maintenance and monitoring of the sensors, decision support and integration into EMR and insulin infusion algorithms,  and very importantly, appropriate staff training, specifically as it relates to the proper placement, calibration, and interpretation of these systems.  Like point-of-care glucometers, CGMs are vulnerable to interferences by substances that can cause readings that are both falsely high and falsely low.  Medications like albuterol, lisinopril, heparin, hydroxyurea, have all been described in the  literature, as well as, very importantly, acetaminophen, which is a very commonly prescribed medication in the hospital setting.  There are newer interference-resistant devices, like the Dexcom G6 shown here, that make this acetaminophen interference a potential thing of the past with normal therapeutic doses.  And I will be discussing more possible interferences and their implications later in this talk. So continuous glucose monitoring systems are an emerging option for us.  Current expert consensus is that this technology is likely both accurate and reliable in the critically ill, but we're still awaiting larger studies in additional populations that  address more patient-centered outcomes. And somewhat fortunately, some of that data might actually be accelerated as a result of the COVID-19 pandemic.  In March of this year, the FDA issued a guidance document that expanded the availability and  capability of non-invasive remote monitoring devices to help better support patient monitoring in the hospital, and this included continuous glucose monitoring systems.  They did this in order to reduce contact and exposure risk for healthcare workers and also to reduce PPE use.  This policy was specifically intended for factory-calibrated systems by Abbott and Dexcom, and this was great, because if you were a system that had access to, excuse me, if you  were a hospital that had access to these systems, it meant that your medical personnel potentially  did not have to enter a patient room, don all the PPE, solely for the purpose of checking a blood glucose level.  And now we're seeing reports in the literature describing patients' hospitals' experiences with this practice, which will help support future clinical practice and use of these  technologies in our ICUs. But in the meantime, we will continue to use one of our workhorse devices in the ICU, our point-of-care glucometers.  These are the current mainstay for monitoring and treating patients with dysglycemia, and happily, they're very convenient and easy to use.  However, many of them do lack accuracy when used in the critically ill, and it wasn't until 2018 that the FDA cleared the first point-of-care blood glucose monitor for capillary  testing in critically ill patients, and this was the Nova Biomedical StatStrip system.  It has also been cleared for venous and arterial whole blood specimens, and it's auto-correcting, meaning that it's resistant to some commonly interfering substances.  It's worth pointing out that not all hospitals use this system in their ICU. It's also important to remember when discussing this that the type of sample matters when  using these systems at the bedside. Plasma is considered our gold standard because it lacks confounding factors like hematocrit and oxidizing substances.  Arterial or venous whole blood is preferred over capillary for patients who are in shock,  who are on vasopressors or prolonged IV insulin infusions, and so the take-home here is really that any questionable results from any of these sample types should be confirmed by  a lab draw. So now that we've discussed a bit about the devices and their sampling types, I want to  talk more specifically about common medication interferences with the devices that you may see at the bedside.  Ascorbic acid has seen a bit of a revival in popularity over the last few years, specifically in sepsis.  Since the early 2000s, we've used high-dose vitamin C infusions to reduce fluid requirements during acute burn resuscitation for large burns.  An early experience with this therapy quickly began to show that it interfered with point-of-glare  point-of-care glucometers, specifically the ones that utilize glucose oxidase enzyme technology.  And this interference was shown to falsely elevate blood glucose readings, which in turn can trigger unnecessary insulin treatments that can potentially cause hypoglycemia.  So this graph was work done in 2014 at my own institution, UC Davis Health in Northern California.  We have a large burn center, and I think it really nicely illustrates how discordant blood glucose results can be between the lab, which is shown here with the black squares, and  different glucometers when using high-dose vitamin C. This interference is now a well-described phenomenon in the literature, and it demonstrates  how important it is for clinicians to be aware of the limitations that exist with the devices that you use at your institution.  So we know that this interference holds true for high-dose vitamin C therapy in burns and  malignancy, but what about the much smaller doses that we're using more recently to attenuate inflammation and vascular injury in sepsis and ARDS?  Several studies have looked at this and have found that serum vitamin C levels are highly dependent on the dose administered as well as on its renal elimination.  The Phase 1 trial from 2014 shown here looked at patients with severe sepsis and demonstrated  that a dosing strategy similar to the 1.5 grams Q6 used in the sepsis literature resulted in plasma levels that were in the 300 micromolar range.  Similar results published in 2018 looked at critically ill patients with multi-organ failure and showed that a 1 gram Q12 bolus dosing resulted in 1-hour plasma concentrations,  again in the 200 to 300 micromolar range. This also showed a linear dose concentration relationship that can be extrapolated to the  every 6-hour dosing regimens that are used in sepsis. So based on this PK data, we can reasonably assume that sepsis dosing for vitamin C is  unlikely to result in plasma concentrations more than 300 micromoles. And indeed, this is shown to be true in clinical practice as well.  I've highlighted two different studies here, one from Pharmacotherapy and another from the Journal of Diabetes Science and Technology, and they both showed very similar results,  demonstrating with their Park's error grid analyses that vitamin C therapy for patients with sepsis is unlikely to cause a clinically significant interference.  So the message here is really that not all point-of-care glucometers are created equal.  The newer meters do have less interference, but not every hospital has updated this technology yet.  So it's important to be familiar with the device at your institution, specifically as it relates to the methodology and potential limitations in accuracy.  For vitamin C therapies that result in very high plasma levels, you will likely see some  interference, so it's best to avoid point-of-care checks in these situations and instead rely on the lab for more accurate results. So what about for CGMs?  As I mentioned before, these are still susceptible to interference, although the newer systems have made it a point to try to minimize this.  Most medications that are used in the inpatient setting have not been evaluated for potential interference, so it's difficult to make a strong recommendation here.  Expert consensus is that current systems cannot replace capillary point-of-care checks in the hospital, but given that we're seeing more hospitals use CGM technology during the  pandemic, I suspect we'll see a lot more in the literature regarding those experiences in the coming year.  Moving on to medications that affect insulin resistance, most commonly of course, the steroids.  Widely prescribed in the ICU for their anti-inflammatory and immunosuppressive properties, they are  the main cause of medication-induced diabetes, and we know that two-thirds of patients receiving high-dose steroids will develop hyperglycemia. So what is the guidance here?  You really have to look at the drug. Stereotypes and pharmacodynamics matter, and I'm not just saying that because I'm a pharmacist. They're important.  They provide the basis for dosing strategies, and understanding them really helps us optimize care.  For example, in the context of hydrocortisone use in septic shock, there was a small prospective  study in 2007 that showed that Ebola's strategy of hydrocortisone significantly increased blood glucose, and this was not seen during the continuous infusion strategy.  This influenced the surviving sepsis guidelines to then suggest IV hydrocortisone at a dose  of 200 milligrams per day in order to minimize the glycemic variability seen with bolus dosing. Outside of septic shock, there's been much research done to try to identify the ideal  insulin regimen for steroid-induced hyperglycemia, and the one I'm showing here is from 2017, and they used a really unique approach to match the two pharmacodynamic properties of  insulin and steroids. They added a specific steroid correction insulin to a basal bolus regimen, and the insulin  amount was actually based on the steroid dose and the patient's weight, and this resulted in a significantly lower mean blood glucose and lower glycemic variability.  This approach is very intriguing, and I love that it matches the PD properties of the drugs, but it does need to be corroborated in a larger study.  It's a little challenging to provide specific recommendations here. There's more research that definitely needs to be done to identify the best regimens for  steroid-induced hyperglycemia, but the basal bolus insulin regimen is the most common strategy,  and we do know that just sliding-scale insulin alone is inferior, so be sure to consider the steroid type and duration of action, and if it's possible to decrease the steroid dose  or duration of treatment, it's best to do that instead of just adding more insulin. The new standards section on glucocorticoid therapy was revised in 2021 to include more  guidance on the use of NPH with steroid-induced hyperglycemia, and NPH might have an advantage  over longer-acting insulin because its effects show a similar timeline to that of the afternoon  blood glucose peaks that we see with steroids, but whether this NPH approach actually improves glycemic control is debated in the literature.  So finally, I'll briefly touch on what I'm calling lurking dextrose in the ICU and how it is related to our efforts with fluid stewardship.  So the discussion about fluid creep in the ICU is relatively new. There were two studies done in 2018 that I unfortunately didn't have time to talk about  today, but they looked at how small volumes of electrolyte repletion and TKO and medication diluents really contribute to the overall total fluid volume of patients in the ICU,  and it was pretty striking. These amounts were not trivial. And here's a real-life example, both of the substantial IV volume that's administered  to our patients as well as of the lurking dextrose, which I've pointed out here with the red arrows and boxes, and these occult dextrose sources can really influence dysglycemia  and glycemic variability. For example, when I did the calculations for this patient, he received a total of 195 grams of dextrose alone from the medication in 24 hours.  So this is a type of iatrogenesis that your pharmacists and dieticians on the team can really help you avoid.  The objective today was to identify meds and other factors that influence glucose control when using diabetes technology, and I have to say there are a lot.  Achieving glucose control in the ICU really requires a strategic approach, and it's important to know why and how glycemic excursions happen before you make the decision to treat.  Continuous glucose monitoring technology is continuously evolving and may be coming soon to an ICU near you, but before that, we need more data for diverse inpatient populations,  more data on what glycemic metrics make sense in the ICU, and more data on the cost-effectiveness of implementing such technology. And that's all I have for today.  Thank you for attending.

Video Summary

The video transcript discusses various aspects of glucose monitoring in the critical care setting. The speakers touch on topics such as the challenges of glycemic control, the use of continuous glucose monitoring (CGM) versus point-of-care meters, medication and other factors that affect glucose monitoring, and the impact of workload on nursing when implementing glycemic control protocols.<br /><br />The speakers mention that CGM can provide valuable information about glucose levels and trends, but it is still being investigated for use in hospitalized patients. Point-of-care meters are commonly used in critical care, but they may have limitations in accuracy, especially in the critically ill.<br /><br />Medication interference is an important consideration, as some medications can cause false readings in glucose monitoring. For example, high-dose vitamin C can interfere with point-of-care meters, leading to falsely elevated glucose readings. CGM systems may also be susceptible to interferences from medications.<br /><br />Steroids, commonly used in critical care, can increase insulin resistance and worsen glycemic control. The choice of insulin regimen and dosing strategies for steroid-induced hyperglycemia should be tailored to the patient's specific medication profile and needs.<br /><br />The video also highlights the potential impact of "lurking dextrose" in the ICU, where small volumes of dextrose from medications and diluents contribute to overall glucose levels and variability.<br /><br />Overall, the speakers emphasize the need for careful consideration of medication profiles, technology limitations, and patient-specific factors when monitoring and managing glucose levels in the critical care setting.

Asset Subtitle

Endocrine, Procedures, 2021

Asset Caption

The Society of Critical Care Medicine's Critical Care Congress features internationally renowned faculty and content sessions highlighting the most up-to-date, evidence-based developments in critical care medicine. This is a presentation from the 2021 Critical Care Congress held virtually from January 31-February 12, 2021.

Learning Objectives

Compare and contrast current technologies for glucose monitoring in the ICU
Examine the nursing burden of glucose monitoring and how it affects management and control of glucose
Identify medications and other factors that influence glucose management and control

Technology in Glucose Monitoring
Michael S. D. Agus

Nursing Burden of Bedside Monitoring and Titration
Ruth Lebet

Tricks and Traps: Other Factors That Affect Your Glucose
Rachelle Firestone

Meta Tag

Content Type Presentation

Knowledge Area Endocrine

Knowledge Area Procedures

Knowledge Level Intermediate

Knowledge Level Advanced

Membership Level Select

Tag Diabetes

Tag Insulin

Tag Point of Care

Year 2021

Keywords

glucose monitoring

critical care

glycemic control

continuous glucose monitoring

point-of-care meters

medication interference

workload impact

insulin resistance

lurking dextrose

	Society of Critical Care Medicine 500 Midway Drive Mount Prospect, IL 60056 USA Phone: +1 847 827-6888 Fax: +1 847 439-7226 Email: support@sccm.org	Contact Us About SCCM Newsroom Advertising & Sponsorship DONATE	MySCCM LearnICU Patients & Families Surviving Sepsis Campaign Critical Care Societies Collaborative	GET OUR NEWSLETTER
© Society of Critical Care Medicine. All rights reserved. \| Privacy Statement \| Terms & Conditions The Society of Critical Care Medicine, SCCM, and Critical Care Congress are registered trademarks of the Society of Critical Care Medicine.