Transcript

MIC CAVAZZINI:                Welcome to [IMJ On-Air] from the Pomegranate studio of the Royal Australasian College of Physicians. I’m Mic Cavazzini and today we’ve got a new face in the pilot’s seat. Dr Mervyn Kyi a consultant endocrinologist joining us from Royal Melbourne Hospital. Please tell us what else we should know about you?

MERVYN KYI:     Thanks for the introduction. I'm an academic endocrinologist and physician in general medicine at the Royal Melbourne Hospital. I'm also an endocrinologist at the Northern Hospital. My academic interest is in acute care in diabetes and endocrinology and having completed my PhD on models of inpatient diabetes care.

MIC CAVAZZINI:                Thank you. And today’s discussion is around a paper published in March for the Internal Medicine journal titled “Heterogeneity in the management of diabetic ketoacidosis in Australia: a national survey.” You happened to be one of the reviewers. What did you think was important about this study?

MERVYN KYI:     Yeah, so DKA is a common and potentially life-threatening complication of diabetes and we have to treat it appropriately to ensure safe patient outcomes. And managing DKA affects many healthcare professionals from the emergency department, intensive care unit to the general wards and it requires close input between medical and nursing and potentially other health professionals such as pharmacists and dieticians. It's also a highly dynamic scenario that requires close monitoring and frequent adjustments to treatment depending on the monitoring. So, it's important to do it right. And it usually follows a protocolized treatment plan. But as we will discuss, there are a lot of variations in treatment protocol which can affect patient outcomes and this paper highlighted that very nicely.

MIC CAVAZZINI:                Okay, let's introduce the authors, one of whom is hosting me in her office at St Vincent's Hospital in Sydney. Dr Lisa Raven, please give us your two line CV.

LISA RAVEN:       Yeah, I'm an endocrinologist at St Vincent's Hospital in Sydney working in the public and private hospitals and doing a PhD at the Garvan Institute of Medical Research.

MIC CAVAZZINI:                And dialling in from Adelaide, we have Dr Mahesh Umapathysivam.

MAHESH UMAPATHYSIVAM:      So, my two-line CV is that I'm an endocrinologist working at the Southern Adelaide Diabetes and Endocrine Service as well as the Royal Adelaide Hospital. I did my PhD at the University of Oxford looking at personalisation of care in type 2 diabetes.

MIC CAVAZZINI:                Mervyn, do you want to take it away with the first bullet point?

MERVYN KYI:     Yeah, so in this paper, you guys wanted to assess the variation in DKA management across Australia. So, you sent the survey questions out to the Endocrine Society of Australia mailing list and the Intensive Care Specialist Treatment Group and received 31 responses from different public and private organisations around the country with an even split between preference for the insulin infusion protocols being used, fixed and variable-rate of insulin infusion. So, did you expect to see such an even split and are the outcomes data out there really so equivocal?

LISA RAVEN:       So, anecdotally we did think that there was variation—we didn't necessarily expect it to be a perfect even split which worked very well for the manuscript. But from our personal experiences and speaking with others we knew that there was definitely going to be variation but we just wanted to quantify that.

MAHESH UMAPATHYSIVAM:      So the lead into the idea behind this study was actually the Hamburg meeting of the European Association of the Study of Diabetes, where for the first time in some time there was an updated position statement from the ADA, EASD, so the American Diabetes Association, the European Association for the Study of Diabetes, as well as a number of other European societies about the management of hyperglycaemic emergencies, particularly diabetic ketoacidosis, for the purpose of our paper, which came down quite hard on their recommendation for a fixed dose infusion and that really prompted discussion amongst myself and Lisa and, I think, Mervyn we discussed this with you later on in the day, about what we were seeing in the Australian setting and variation even within our own institution about what was happening despite management protocols. So, there was, I guess, this rationale to work out what exactly was happening in the Australian community.

MIC CAVAZZINI:                Speaking as the non-clinician here, I was surprised to see there were differences also in the finer details of the protocols. One of the fixed-rate protocols had a set starting rate of 4 units per hours, with the others determined the rate based on units per kilogram. There were even slight variations whether it’s  0.1 unit per kilogram 0.05, 0.07. And then 3 of the 14 variable infusion protocols also included a bolus of intravenous insulin at the start of DKA management. So, if we can’t say whether fixed or variable is better, can you  even speculate on these finer differences?

LISA RAVEN:       Yeah, I was quite surprised by some of that variation within the groups and I guess the recommendations from this joint consortium that was presented at the meeting in Hamburg was for fixed infusion rate of 0.1 units per kilo and then it does actually recommend a reduction down to 0.05 units per kilogram. But the fact that there's so much variation even within whether you do a variable or fixed-rate insulin infusion just goes to show that everywhere does do it slightly differently, so I was quite surprised by that variation within the groups.

MAHESH UMAPATHYSIVAM:      I think it's also probably important to note that there are similarities. There seems to be sort of key pillars of diabetic ketoacidosis management which are reflected in all the protocols. So, it's clear that IV insulin administration is important as part of an infusion, correcting the fluid deficit with hydration, and then close electrolyte management is key. And the protocols reflect that, as well as a focus on addressing the underlying precipitant. I guess the thing that varies, and probably as a reflection of the lack of strong robust evidence, is how the insulin and fluids are delivered between sites and that's really evolved I think over time as consensus and expert opinion has changed but not really driven unfortunately by evidence. And it's difficult to answer those questions. It requires a lot of effort to address each of those individual questions in a progressive way.

MERVYN KYI:     Yeah, so we were also surprised that two thirds of the Australian protocols actually express a specific resolution point for DKA. For some the pH threshold being 7.3. But there seems to be a fair degree of variability in the ketone threshold that defined resolution of ketoacidosis. And that vary between 1 millimoles, 0.6 millimoles, even 0.3 millimoles per litre. So, I guess, can you comment on why there are such differences and is it really important and or is it more of the trajectory that's important?

MAHESH UMAPATHYSIVAM:      Yeah, so we did see that there was differences in the ketone cutoff and they were all at the lower end of the spectrum, so one or below millimoles per litre in terms of beta hydroxybutyrate and the vast majority falling at 0.6. So, I guess when we're thinking about resolution, to my mind probably what's important is what—we're talking about transition from intravenous to subcutaneous insulin and how do we do that safely without the concern about rebound ketosis or ketoacidosis. And, I guess, also minimizing the amount of burden of care with an intravenous infusion as well as the associated complications that come with that.

I don't think we know what the safe ketone concentration is for transition from IV to subcutaneous insulin infusion. And I think further studies are required to work out if there is increased glycaemic variability if we say come off an insulin infusion at 1 millimole ketones versus coming off at 0.3. 0.6 seems to be a sort of a physiological sort of threshold that—well it's certainly recommended by the ADA position statement but whether or not it makes a big difference, whether it's 0.8, 0.6, I don't think that's going to be a huge determinant. Probably other clinical factors are probably also going be a major driver, so if your patient's looking well, eating and drinking, and has already potentially started some of the transition to subcutaneous insulin, it may be appropriate to determine that their DKA has essentially resolved, and then transition to subcutaneous insulin.

LISA RAVEN:       Yeah, I think that the fact that there is such varied cutoffs does reflect clinical practice and the realities of providing intensive care for these patients. Often, once the ketone level is below one millimole per litre, people are lot more comfortable. And most of these protocols, not captured within our survey, but requiring hourly checks of glucose levels, ketones, potassium and so the frequency of monitoring may become less after that point, that could be also a driver of not having a specific cutoff. And if a cutoff is mentioned and someone is very well but their ketone level is 0.7 millimoles per litre then how long are you going to keep doing those checks might come into the factor as well.

MAHESH UMAPATHYSIVAM:      I think that's an important point Lisa because we're all working stretched healthcare utilization. And certainly, one of the other big findings was the variations in location of care. So, keeping a patient in the intensive care unit to continue to administer an insulin infusion can be very burdensome to a healthcare setting and when those precious beds are being sought after by other patients who are much sicker. There is a pressure to make sure that your patient gets the appropriate setting of care and if a rigid 0.6 cutoff might not reflect the actual clinical setting.

MIC CAVAZZINI:                I guess it's like any conversation about guidelines and algorithms, they help the less experienced staff more than they help senior staff who can judge with a bit of a gestalt kind of how well the patient is doing, they don’t need to be bound as tightly by the precise cutoffs...

LISA RAVEN:       Yeah, that raises a good point that wasn't captured in the survey but—but deviations from protocol and how frequent they can be as well and often that's a senior clinician making that decision. And they are there as a loose guide to try and help the system work better.

MAHESH UMAPATHYSIVAM:      And I think one of the points Mervyn made in the introduction about how rapidly the physiology of DKA changes with treatment progression through treatment. In the majority of healthcare settings around Australia, you're probably not getting your senior person who's making all those decisions. And having guidelines which allow the application of the knowledge and the expert opinion sort of throughout the process of DKA is really important. We've seen the reduction in mortality with introduction of guidelines, like, the mortality in developed countries is less than 1 per cent on using these protocols. I mean, they're highly effective. And it is a difficult space because of the rapidly changing physiology and it gives sort of the junior doctors who are probably going to be doing the majority of the—and the nursing staff at the bedside—a good framework to work within.

MERVYN KYI:     Yeah, so I guess a little bit of a digression from the current paper. But the other thing to highlight that I've noticed in clinical practice—we've also published a bit on this—is how common ketosis is these days. So, you know, in the past, we used to do ketone testing when glucose is above 15. And then it became, if you're on the SGLT inhibitor, you probably should do a ketone test even though they're not hypoglycaemic because of the euglycaemic ketoacidosis. And that's led to a lot more ketone testing. And I think we found that a lot more people have mild degrees of ketosis that may or may not be clinically relevant. And so, this 0.6 magic number has been there for a long time, but it doesn't necessarily reflect the current level of practice. And perhaps it goes back to your paper in that perhaps as the environment in our clinical practice changes, perhaps some of our guidelines may need to be updated to reflect some of these changes.

LISA RAVEN:       Yeah, we did specifically ask about whether a different protocol is specified for SGLT2 inhibitor-associated ketoacidosis. And most of the protocols didn't have a specification about that yet, but I suspect in coming years that will also be an important topic. And I think Mahesh has done some research on that separately to this paper.

MAHESH UMAPATHYSIVAM: The research that Lisa's alluding to is—one of the observations from a retrospective cohort that we have of patients treated with SGLT2-associated ketoacidosis and the observation that dynamic infusions which were used by the South Australian metropolitan networks really performed much worse than the traditional type 1 cohort. And that's probably because of a tendency towards euglycemia and with the premise of a variable dose or dynamic insulin infusion is that we're leveraging the association between ketosis, or ketone production, and high blood sugar to determine the insulin dose. And when that link is less closely associated with SGLT2 inhibitor use, where you have a tendency to euglycemia and ketosis, then the insulin infusion may not be adequate. And so, we saw in that cohort that there was almost a doubling in the time to resolution of age-matched patients with SGLT2 associated DKA as opposed to type 1-associated DKA. So, I think an appreciation that there's a difference in response to SGLT2-associated DKA on some of these infusions is important. And then the flip side to that is if you're using a fixed dose infusion, the risk of hypoglycaemia is probably much greater in the SGLT-2 group. So, early introduction of dextrose-containing fluids I think is important in that setting.

MIC CAVAZZINI:                Do we need to explain for non-endocrinologists like me how you get ketosis at normal glucose levels? Is fat metabolism triggered by the drug directly?

MAHESH UMAPATHYSIVAM:      So, it's not fully understood, but I think the broad strokes of the mechanism are relatively clear. So, we know that ketosis develops when there is an increase in the breakdown of fatty acids, which is driven by, either in the case of type 1 diabetes, an absence of insulin or relative reduction in insulin and an increase in glucagon. So, that ratio of glucagon to insulin is important. And in SGLT2-associated DKA, what we see is urinary loss of glucose, lowering blood sugar, reducing the stimulation of insulin secretion, but also increasing glucagon secretion. So, you achieve that ratio of elevated glucagon to insulin through a different mechanism. But the result is still ketone body formation and then ketosis and progressive acidosis. But it occurs, at times, a normal blood sugar.

MERVYN KYI: And I guess the other thing to add to that is that with the SGLT2 inhibitor-related DKA, often the effects of the drugs are long-lasting—you know, the half-life is anywhere between 12 to 24 hours—so, occasionally people can continue to have ketoacidosis or at least ketosis for a period of time where some of our protocols, which are all based on hyperglycaemia management, in conjunction with ketosis management, doesn't really doesn't really resolve the ketosis or the DKA as effectively. And we need to have much more prolonged insulin infusions in that scenario while the drug's on board.

MIC CAVAZZINI:                Okay, let's stop nerding out and get to the crux of the paper, which is what are the costs of this variability? Three quarters of survey respondents said they'd worked at another hospital that had different DKA management protocols. How much of a burden is that on doctors and nurses moving between institutions? Is that potentially one of those holes in the Swiss cheese that we should worry about from a systems and safety point of view?

LISA RAVEN:       I think the fact that we can't show which of the protocol types is superior, based on the limited evidence that's out there, I think that actually having exposure to different protocols might help in aiding understanding of the mechanisms and therefore almost learn a little bit more rather than just purely following a protocol blindly. But I guess, for the nursing staff I do think it's really hard for them. Particularly as they are following the protocol to know where whether to titrate insulin or fluids and one of the limitations and concerns that I do have over the fixed-rate insulin infusions is that that decision is to titrate the type of fluid, and that's a pretty big decision because you need to know a lot about their comorbidities, renal function, cardiac function, to know how much fluid someone can actually handle. So, that's my concern personally.

MAHESH UMAPATHYSIVAM:      I think sort of cognitive burden of junior doctors is important, moving between institutions. There is a cost associated with that. I think there's also a potential safety risk, as Lisa mentioned, when patients are transferring between sites and Australia, unfortunately, a lot of our care is spread across large distances. And where patients are transferring from one institution, usually a lower care setting up to a higher care setting, moving from a fixed dose to a dynamic or dynamic to fixed is probably a high risk setting for one, loss of efficacy in terms of ketone clearance and resolution of DKA, but also hypoglycaemia risk. So, insulin is a very high-risk medication and protocol variations between hospitals, I think adds an additional layer to that risk.

MERVYN KYI:     So guys, I guess. despite the lack answer as to which regimen gives better outcomes, there’s been a flurry of guidelines in recent years encouraging the use of fixed-rate insulin infusion, like you mentioned; the ADA,  the AACE and then the ESD and the Joint British Society for Inpatient Care. The fixed-rate insulin has been recommended despite perhaps potentially higher rates of hypoglycaemia with that protocol. So, do you think that all those consensus guidelines are—they’re intention is to try to simplify and standardize and could that itself bring  greater benefits in actually standardizing things so that when people move around they're using the same protocol? And do you think that's better than our current practice in Australia?

MAHESH UMAPATHYSIVAM: Yeah, so I think there are two things to consider. So, I think the ADA and EASD position statements do a fantastic job of trying to simplify management. There are clearly three pillars that they focus on in terms of the management, the fluid resuscitation, insulin administration and electrolyte management—they're key. And certainly, there are big gains to be had by having a simple DKA protocol. And I think that was one of their key focuses; was that a junior doctor or nurse can pull this up at 2am in the morning when this patient's presented and start high-quality care.

But I think in terms of delivering care, the other flip side to that is understanding local factors which affect management. If you have many years of institutional knowledge, around how to run a dynamic or variable dose infusion, I think suddenly flicking to a fixed dose infusion is probably going to outweigh some of those benefits of simplicity. And this is sort of anecdotal, I don’t think this has been specifically looked at—I do feel that the fixed dose infusions probably are slightly more intensive in terms of requirements from a nursing side of things. If you don't have the ability to meet those needs in your care setting, then rigidly applying a fixed dose infusion might not be appropriate. And so, some of the simplicity and the rigidness of that approach might not lead to improvements in care. So, I think simplicity, yes, is hugely important, but understanding of local factors is probably going to potentially outweigh that.

LISA RAVEN: Yeah, think having a protocol is definitely crucial they had to pick one. I did do a little bit of a deep dive to try and understand that and all I could find was that the Joint British Diabetes Societies for Inpatient Care Group, and they are the only ones who've listed anything as a rationale. And they had said it was due to patient demographics with increasing obesity and insulin resistance, I guess what they're saying is because when they talk about fixed-rate insulin infusions, it is weight-based. They're talking about the point one units per kilo. And in general, people will receive a higher rate of insulin on the fixed-rate insulin protocols. Because say someone who's 70 kilos is going to be receiving seven units, which is quite a lot of insulin, whereas depending on what their glucose level was, they might only receive five units. So that was their rationale, as well as the need to have a more consistent protocol around the country.

The more recent guidelines do actually have a very small caveat about the use of variable insulin infusions for nurse-led centres, which I guess are there almost acknowledging what Mahesh was saying that potentially it is more labour intensive. But then it's almost that that recommendation is limited to nurse-led centres, whereas really, I don't see a reason why it can't be in other centres. And the other interesting point was I recently went to the Endocrine Society meeting in America and there were some presentations from American hospitals who using AI algorithms to help them with insulin titration, and that's definitely a variable insulin infusion rate that they're using. So, I think that there are places around the world that are still using variable insulin infusion rates despite this global recommendation.

MERVYN KYI:     Mic, I think the guys have done a really good job of pre-empting all our questions. They've already answered most of our questions, which is fantastic.

MIC CAVAZZINI:                Absolutely. Very smooth. Is it worth going into that Australian study led by Dr Anojian Koneshamoorthy? Should we be guided more by local research conducted in local settings? Other factors in Australian settings, you know, preferences for fluid or other factors that would steer you in one direction or the other that might not map onto the British experience?

MAHESH UMAPATHYSIVAM:      Yeah, I think there are two probably studies that are relevant. Lisa, you were an author on one of those papers, I believe? So why don't you, I don't want to steal your thunder. You can describe your own study.

LISA RAVEN:       Yes, so I was a co-author on a study based out of Canberra. It was a small retrospective study,  and that was looking at a changeover period where a hospital had gone from a variable-rate insulin infusion to a fixed-rate insulin infusion. And some interesting points from it were that having this resolution of DKA is very useful both clinically as well as potentially for looking into research and finding how often a ketone level within a certain cutoff has been measured. Also, it really demonstrated that issue that we discussed before about particularly more regional centres having to transfer patients to a more metropolitan centre for escalation of therapy. And even just going through the available data, it was very difficult when someone had been started on an insulin infusion at a different hospital and then knowing exactly what the protocol was and even just that travel time, what was happening to people's glucose levels and fluids and insulin infusion during a transfer. Because, in Australia, how far people do have to travel and also time for staff at different hospitals to get used to new protocols. These are all things that had to be considered. But I do think that more Australian data is required because of our unique settings and I think hypoglycaemia, like certainly in New South Wales, there are very strict financial implications of hypoglycaemia. So, all hospitals are monitored on how often hypoglycaemia occurs and there has been research, albeit mainly retrospective, that the fixed-rate insulin infusion protocols do have higher rates of hypoglycaemia. So, I think that's important for an Australian setting.

MERVYN KYI:     Perhaps if I can also add, again, in clinical experience and perhaps going back to the British guidelines, as I understand, and I've spoken to some of my British colleagues, there are many centres which has a medical HDU or medical high-dependency unit where if a certain particular criteria are met on people presenting with DKA, they might go towards a medical high-dependency unit where perhaps giving a fixed-rate of insulin infusion is appropriate because there is slightly more intensive nursing input availability, and perhaps they are able to increase the dextrose infusion as the glucose drops. Whereas, I think in Australia, particularly in Melbourne, there aren't too many medical HDUs. They either go to the ward or they go to ICU and sometimes patients are not that unwell to be in ICU. They can be in the ward setting, but in that scenario, I think it is not that safe giving a fixed-rate of insulin infusion where the dextrose infusion has to be ramped up as the glucose drops tremendously. So, I guess I agree that at least in according to the local resources and local procedures, we probably should tailor our infusion protocols to what suits us.

MIC CAVAZZINI:                Excellent, thank you, Mervyn.

MAHESH UMAPATHYSIVAM:      Yeah, I think there's one other point if I can add to that, think, and Lisa brought it up initially. With the fixed dose infusions, if you're adjusting fluid rates to maintain blood sugar, you're tying the volume of fluid that's going in to a patient's blood sugar. And increasingly we're seeing patients with type 1 diabetes living longer, becoming more comorbid particularly from a heart failure point of view and that does, I guess, add an additional element of risk in terms of exasperating heart failure. And then particularly with the indications for SGLT2 inhibitors, where you're getting a patient population who have an indication for renal dysfunction and for heart failure—having a fixed infusion where you're not taking into account the fluid status but rather the glucose is a driver of your fluid resuscitation, you may run into potential harm associated with that.

MIC CAVAZZINI:                Mervyn, did you want to go into the performance indicators?

MERVYN KYI:     Yeah, I guess we more or less covered it, but I guess I'm just going to ask that again, guys, just to perhaps have almost a conclusion statement to the findings of the study. But I guess, the question is, ultimately, does it really matter, right? Whether or not we follow a fixed-rate of insulin infusion or dynamic or variable-rate of insulin infusion as long as the patient is getting the appropriate pillars of therapy, including intravenous infusion, fluid resuscitation, electrolyte replacement, et cetera. So, on one side, there are some metrics such as whether or not a person can have ketoacidosis clearing quicker and therefore perhaps shorter length of stay. Versus the other metrics in potential harms such as hypoglycaemia or hypokalaemia during intravenous insulin infusion. And I guess, know, perhaps your overall thoughts as to whether we should all move to one type of infusion, would you be open for us, for everyone to have different types of insulin infusion to suit their needs? And if you had to pick, would you choose one or the other?

MAHESH UMAPATHYSIVAM:      So, I think that we should endeavour to achieve both fast rates of resolution and avoidance of those key complications that you manage, hypoglycaemia, hypokalaemia. And I think that we can achieve that in a protocolised way if we have the appropriate in-built monitoring as well as replacement regimens from a fluid and electrolyte point of view. And I don't think necessarily that the make or break of that desire to get those attributes is going to be driven by whether or not you use a fixed or dynamic infusion. I think there's inadequate evidence, at the moment, to recommend one over another. What I think is important is you have a protocol that works for your local environment. And if you have extensive institutional experience with one over another or local factors which make one more appropriate than another, then I think that's what you should be doing in your institution. The flip side of that is we have to accept a degree of variability in management across Australia until there's sufficient evidence to recommend one over another in terms of either the efficacy or safety of the infusion.

MERVYN KYI:     I guess this is exactly the type of question that should be asked of the upcoming Australia National Impatient Diabetes Audit and Survey that we'll be conducting, Mahesh and I are both part of that subcommittee from the Australian Diabetes Society to convene a nationwide audit, similar to what the UK has done 15 years ago. I would suggest t we definitely should be asking and potentially asking to standardize some of these criteria.

LISA RAVEN:       Yeah, I completely agree. I think that in an ideal world, the fact that we do have such variability in Australia potentially lends us the opportunity to have these comparisons and contribute to the data. We did try to put together a national retrospective survey or even prospective, however, it is challenging with so many different research, ethics, offices, different barriers across Australia. But at least if we could have some sort of agreeance on the resolution of DKA, then we could at least try to compare and contribute. It will be a challenge to compare in some ways because some hospitals mainly manage DKA in intensive care units and therefore that's quite a different setting with different opportunities, whereas some other hospitals, the staff on the wards are much more comfortable. So, it will be very interesting to see from that side of things. One of the things that did come out of the survey was that the intensive care unit locations had higher rates of variable, which isn't what we would have expected, as we talked about, potentially an increased workload. So, the wards were very split and high dependency units tended to have higher numbers of fixed-rates, insulin infusions. In terms of the states, there was a little bit of a trend, but again it's very small numbers so it's hard to make too much of this.

MAHESH UMAPATHYSIVAM:      So South Australia, I think exclusively, used dynamic infusions amongst the endocrinologists. I'm not sure historically why that's the case. And I would be hopeful that given that there is, I think, based on our discussions today, a degree of clinical equipoise, there'll be opportunities for prospective randomized controlled trials that will give us much greater information about a fixed versus dynamic infusion. But then the question is always asked, like we've got a recommendation for the best fixed dose infusion, but there's variation even within the dynamic or variable infusion, what we choose as the comparator.

MIC CAVAZZINI: And that sounds like a very tricky intervention study to conduct in the one setting. Suddenly, you've got how many teams of performing one or the other, obviously not in a blinded fashion, but is that what's held back, the resolution of this question?

MAHESH UMAPATHYSIVAM: I mean the treatment is so effective, right? It's hard to—I mean, it's one of the few conditions where we see resolution within the time of day that you're at work, right? It's hard to say the treatment isn't already very good. We're really aiming for sort of that last 5 percent, that was an arbitrary number, but to optimize care, I guess it is difficult because these patients are coming in very unwell. They may not be in a position to consent. You have to catch them very early, otherwise the standard treatment—if you're two hours late to recruit the patient, there may well be majority resolved. It's such a fast evolution in response to treatment that it's, I guess, something that makes recruitment very difficult. And you're recruiting patients in a time when they're vulnerable, maybe having impaired conscious state. So, I think that's the main challenges around DKA management, not that it's not doable, I think there is the opportunity to do it. But I think it would be very hard to sell to someone. We might reduce your time to resolution by half an hour. Can you please change all your institutional protocols around the management of DKA? I think that's difficult.

MERVYN KYI:     I suspect we will not get level A evidence for this.

MIC CAVAZZINI:                Yeah. Dr Rahul Barmanray who's appeared on the podcast a couple of times. He threw in some interesting references for me to look at. One was the UK’s DEKODE model. It’s a registry of sorts to monitor adherence to the British protocols and prompt training or awareness where that was falling. The interesting thing is the intervention did improve better adherence but there was no significant improvement in the median time to DKA resolution or length of stay across participating hospitals by the end of the study timeframe. Is that disappointing? It sounds from what you've been saying, that it isn't terribly surprising,?

LISA RAVEN:       Yeah, think probably, you know, yeah, the differences are probably not that great either way. I think monitoring of sticking to a protocol, it does make me little bit concerned because as we mentioned before about deviations from protocol in a specific setting. If you're getting sort of flagged, if it's for a clinical need, then is that really helping us or the patient? But I mean, the NHS has some flaws but it is very good at this sort of data collection and the fact that they did all just you know go to one type of insulin protocol to allow them to have this sort of framework. And I guess that also comes back to one of Mervyn's questions from before of, does this really matter? I suspect the real main messages are quick initiation of therapy with intravenous insulin, IV fluids and electrolyte replacement. That is the main thing. Mahesh might have some thoughts from his time in the UK.         

MAHESH UMAPATHYSIVAM: Not so much the time in the UK. But I do have some thoughts about the introduction of electronic medical records and the ability to audit our practice. And I think we've seen this from some of Mervyn's studies about the role of inpatient diabetes teams and sort of consistency of care and proactive involvement and some of the quite large benefits, particularly in the realms of hospital-acquired infection. The introduction of widespread use of electronic medical records I think will allow us to look between institutions at how we perform in terms of time to resolution which is biochemical. And they get very robustly and objectively recorded in the electronic medical record. Similarly, the outcomes which are undesirable, hypoglycaemia, hypokalaemia, again objective, very well coded. There is the opportunity to compare rates of these things and if you see outliers in care, they provide both an opportunity for identifying what's being done well, as well as areas for improvement. I guess thinking about—so I'm based in Adelaide—in the South Australian context, we're lucky in that all of our public hospitals use the same electronic medical record and the enterprise wide, so all of say health initiatives sort of around safety and quality and governance might give us that opportunity to say, is there deviation from the general sort of rates of resolution and those complications and does that give us any insights into which protocols? Because there are variations in care within public hospitals within South Australia.

LISA RAVEN:       Yes, we're very jealous from New South Wales about that linked system, but we are hopefully moving to a single digital patient record. hopefully we can do that in a few years time as well.

MIC CAVAZZINI:                Any parting thoughts, Mervyn?

MERVYN KYI: No look, I think you guys did a really, really good job in highlighting this issue. I definitely agree with your discussion points in the conclusion statements in that I don't think that we are ready to, for everyone to just go straight to a fixed-rate of insulin infusion, despite all the international guidelines and consensus statements. I think that we should allow for local variations depending on what is available and what is safe for our patients. So, I think that's probably the right thing and we may never get a level one evidence for which particular infusion is better. But on the other hand, I think you're right. And perhaps we should highlight that even more that there was a fair degree of consensus in making sure that the pillars of management were done appropriately and that most of the time our patients do get better and we do deliver a pretty good level of care across the board.

MIC CAVAZZINI:                                Thank you all, very well done.

MAHESH UMAPATHYSIVAM:      Thank you, Mic.

MIC CAVAZZINI:                                Really interesting.

LISA RAVEN:       Well it’s actually just like us talking back when this all started.        

MIC CAVAZZINI:                Many thanks so much to Lisa Raven, Mahesh Umapathysivam and Mervyn Kyi for contributing their time and expertise to this episode of [IMJ On-Air]. You’ll find a link to the article in the show notes at racp.edu.au/podcast. It’s titled “Heterogeneity in the management of diabetic ketoacidosis in Australia: a national survey” and published by the Internal Medicine Journal. The College also boasts the Journal of Paediatrics and Child Health and the Occupational Medicine Journal, which are edited by the generosity and commitment of members.

Just as this podcast is. I really appreciate doctors Aidan Tan, Hugh Murray, Stephen Bacchi and Aafreen Khalid for taking a listen so it could be polished up for your ears. If you have an article you’d like to put forward for discussion in one of our or [Journal Club] or [Case Report] formats, please get in touch via the email podcast@racp.edu.au. I’m always keen to know what you’d like to hear more or less of.

If you’re really enthusiastic, you can even leave a review at Apple Podcasts, Spotify, Castbox or any other podcasting app. Pomegranate Health is now also streamable from the YouTube channel which has the handle @TheRACP or web address youtube.com/user/RACP1938. And finally, you can also subscribe to email alerts for new episodes right from the website. This podcast was produced on the lands of the Gadigal clans of the Yura nation. I pay respect to their elders past and present. I’m Mic Cavazzini. Thanks for listening.