Transcript

MIC CAVAZZINI: You’re listening to Pomegranate Health, a podcast for physicians of the RACP.

I’m Mic Cavazzini, and this episode is about pharmacological management of Type 2 diabetes. We’ll start by focusing on a polemic around exact blood sugar targets that arose this year between the American College of Physicians and several other diabetes organisations around the world. It all comes down to one’s interpretation of just four key trials, and as we’ll hear today, the devil is in the detail. The debate also needs to be reframed in the context of new drug classes that appear to treat some complications of diabetes directly.

Before we get stuck in, let’s refresh some basics principles. The haemoglobin A1c assay is like a long—exposure photograph of the circulating glucose picked up by red blood cells over the last few months. The results of this test are reported as the proportion of glycated haemoglobin over total haemoglobin, and a healthy range is between 4 and 5.7 per cent. These days, clinical labs prefer to express these values as 20 to 39mmol per mol.

When blood glucose remains too high over a long time, the metabolic stress on blood vessels has serious consequences. On the macrovascular side, there’s an increased risk of myocardial infarction and stroke. And the microvascular complications can manifest as vision loss, kidney disease, and sensory numbness.

For two decades, the mantra for blood glucose management has been “the lower the better.” This approach was first informed by a 1993 trial in patients with Type 1 diabetes. It showed that patients treated with multiple doses of insulin every day for six years experienced fewer microvascular complications than those given just one or two doses a day.

Five years later, the UKPDS trial made similar findings in patients recently diagnosed with Type 2 Diabetes. Specifically, it showed that drug therapy titrated to an HbA1c target of seven per cent provided some protection against retinopathy. A follow—up of these patients ten years later also found that there were fewer heart attacks and deaths in that group. This prompted three further studies intended to explore whether more intensive targets of 6.5 per cent or lower could produce similar results in patients already suffering long—term complications of diabetes. The 2008 ADVANCE trial and 2009 Veterans Affairs Diabetes Trial showed some improvements for nephropathy from tight glycemic control, but no macrovascular benefit. Another study called ACCORD did demonstrate reduced risk of non—fatal heart attack, but the trial had to be stopped early because of large increase in all—cause mortality.

To make sense of all these findings, I spoke to endocrinologist Dr Paul Drury from the University of Auckland, who wrote an excellent synopsis for the Better Practice Advisory Council, and has contributed to much of the guidance material in New Zealand over the years.

PAUL DRURY: Hello, I’m Paul Drury, until recently Clinical Director of Diabetes Services for the Central Auckland DHB. I’m still active academically and I work one day a week for the Ministry of Health as Clinical Advisor in Diabetes.

MIC CAVAZZINI: Let’s start by looking at the macrovascular complications. In defending the American college guidelines, the ACP President argued in a press release that none of those three more recent trials have showed a reduction in all—cause cardiovascular mortality. This is supported by a couple of meta—analyses, despite a nine per cent reduced risk of major cardiovascular events compared to the normal treatment group. What do you make of this apparent contradiction?

PAUL DRURY: I think, even though you’ve got 27,000 patients your numbers of events is not absolutely colossal—then some of these things are going to be really, really hard to show. The other issue is that we are getting a cohort of patients with very advanced vascular disease and it may well be that if you delay a cardiovascular complication then a renal complication will turn up in six or nine months. The other issue is if there is a specific disadvantage of tight glycaemic control in that situation then that might counterbalance what would otherwise be a benefit, but I have to say, I don’t think I know the answer to that.

MIC CAVAZZINI: Now let’s consider the microvascular outcomes in each of the trials. In ADVANCE it appears that tight glycaemic control reduced the incidence of microalbuminuria, a surrogate marker for kidney disease. In the Veterans Affairs trial too, there was a reduced rate of progression of microalbuminuria over the treatment period. In the ACCORD trial the intensive treatment group showed less neuropathy as measured by sensation to light touch and the ankle jerk reflex. And the main result in the UKPDS study was reduced need for retinal photocoagulation. How do you synthesise such disparate findings from those studies?

PAUL DRURY: I find it very difficult to synthesise them, and if you look at the studies you’ll see that the sort of candidates, patient groups for each are subtly different—the outcome markers are subtly, or not—so—subtly different. And that I think, to some extent, we’ve ended up with a situation where the meta—analyses are not giving us useful, new information. I’m not meta—analysis expert, so I’d better not sort of tread any further down that line. You know, we need to have high quality studies, big enough on individual or just a couple of questions in order to take things forward. But the problem with all these—in order to show these sorts of differences, you have to have really massive numbers, and you have to have the same assay in the same place— all of which is difficult to do.

***

MIC CAVAZZINI: In April this year, the American College of Physicians published new guidelines stating that a rather relaxed HbA1c range of 7 to 8 per cent (or 64 mmol per mol) was fine for “most patients with type 2 diabetes”. Even more controversially, the next statement reads that "Clinicians should consider de—intensifying pharmacologic therapy in patients who achieve HbA1c levels less than 6.5%.” They pointed to the worrying results from ACCORD, as well as the complications of hypoglycaemia such as disorientation, loss of coordination, irregular heart beat, and loss of consciousness.

The ACP are not the only skeptics about an intensive approach to glycaemic control. In a JAMA Clinical Reviews podcast the presenter Professor Ed Livinsgston argued that UKPDS has had a disproportionate influence on practice. The main claim of the study was that lower HbA1c targets were associated with a 12 per cent reduction of any diabetes—related end—point averaged from 21 different measures. But most of the effect was driven by just one surrogate marker of retinopathy. Meanwhile a 2016 review of advisory statements and consensus guidelines published in the preceding ten years suggested that most of them overplayed the cardiovascular benefits, despite the equivocal evidence.

But the release of new targets by the American College of Physicians prompted an immediate backlash from other experts. The President of the American Association of Clinical Endocrinologists said in a statement: "This is absolutely wrong and regressive—the ACP has stepped back in time.” Others bodies in the US, Canada, the UK and Australia wrote an editorial in the journal Diabetes Care stating that “the ACP recommendations fail to consider several important bodies of scientific evidence—if they are widely adopted in clinical practice, the recent progress in management of diabetes may be threatened.”

One of the authors of that paper was outgoing President of the Australian Diabetes Society, Professor Sophia Zoungas, who was also lead author on the 2014 follow up of the ADVANCE trial. She’s a senior staff specialist at Monash Health and heads the Metabolism, Ageing and Genomics division at the university.

SOPHIA ZOUNGAS: I assume that one considers retinopathy, nephropathy and neuropathy as secondary endpoints in a trial, but in reality these are really clinically significant events. To my mind there’s no question that more intensive glycaemic control reduces the renal outcomes. And indeed in the ADVANCE trial that was Australian—led, we actually showed that there was a reduction in risk of end—stage kidney disease which is, you know, a hard—core renal outcome, and that included requirement for dialysis and renal death. The only thing that seems to be a little different is any benefit for neuropathy, and I think that’s just the fact the way we measure it in these trials. So, I don’t think they’re incongruent from each other. I think it’s just got to do with the background risk profile of the populations being studied differing. And in ADVANCE, we also showed that there was legacy benefit; that those effects continued when the participants were followed for an additional five years.

MIC CAVAZZINI: Is this the explanation—I mean the UKPDS seemed to show very strong effects, but those patients were quite young, the average age was 54 and they tended to be newly diagnosed. As compared to the other three less dramatic studies where the average ages were 60 to 66 and the patients had been ill for at least eight years in those studies. Can you explain the rationale or the presumed effect?

SOPHIA ZOUNGAS: So, there’s a concept of metabolic memory or legacy effect. What we’re describing is that a period of good glycaemic control early in the disease course, where there may not necessarily be end—organ or end—tissue damage, subsequently changes the trajectory of risk for that individual such that they have better long—term outcomes. And you know there’s thoughts that it’s related to cross—linking or advanced glycation end—product production within the vasculature. You know there’s a lot of things that have been postulated but clearly, it’s about getting in early before the cardiovascular disease or atherosclerotic disease is established.

MIC CAVAZZINI: So, if you just look at the ACCORD trial—as we mentioned it was stopped early because of the increased mortality in their intensive treatment group—it’s been suggested that the increased mortality was not in patients who reached the 6.5 per cent target but in patients who started from the highest baseline or the rapid rate at which A1c was brought down. Has there been any clearer understanding of what would explain those findings?

SOPHIA ZOUNGAS: Indeed, it was those people that were being intensified and couldn’t achieve a lower target that experienced the more adverse outcomes. My personal view is that you know these were people that a significant proportion already were on insulin and then they added further therapy or titrated the insulin very aggressively within a sort of window of three to six months. Generally, in clinical practice that isn’t what we do— we’re much more conservative and modest. Indeed, actually in ADVANCE that much more conservative approach was not associated with any increased mortality.

MIC CAVAZZINI: Your co-author William Cefalu asks clinicians to imagine a patient taking metformin who might be given an add—on therapy to deal with an existing cardiovascular risk factor. But if that were to bump HbA1c down to below 6.5 per cent, the ACP guidelines would be to instruct you to remove this potentially life—saving therapy.

SOPHIA ZOUNGAS: So in fact, now in this contemporary period is the first time that we have therapies, traditionally known for glucose lowering, that indeed are showing that they do have independent cardiovascular benefits. And as you suggest now we’re experiencing this situation where we would want to add that second cardiovascular therapy, and if we based the decision on a glycemic target would not do so. But if we based it on a decision of cardiovascular benefit then we would.

You know I saw someone recently who all of their therapy had been stopped and they presented with diabetic ketoacidosis, and they were elderly Type 2 apparently and nearly died of the diabetic ketoacidosis. So, there’s a risk of over—treatment but there’s a risk of under—treatment that goes beyond long—term vascular complications.

***

MIC CAVAZZINI: It goes without saying that diabetes care is not just about drugs. Just as important are counseling around diet, physical activity and smoking cessation as well as a network of supportive health professionals and carers. We don’t have time to go into all of these angles— but the idea of personalized therapy extends to drug regimens too. The 2009 position statement from the Australian Diabetes Society is a great example— and clearly navigates some of the controversy we talked about earlier. Glycated haemoglobin targets should be appropriate to an individual’s disease history, their age and metabolic profile as well as social or behavioral factors.

So in people who are recently diagnosed with Type 2 diabetes and otherwise healthy you can go for gold and aim to get below 48mmol/mol or 6.5%, with the hope of emulating outcomes from UKPDS. Given the many possible barriers to this, however, the Australian guidelines allow that “For most people a target up to 53mmol/mol or 7per centis achievable.” Finally, they say, “In those who do have major co—morbidities or reduced hypoglycaemia awareness, then it is appropriate to shift the target to 64 mmol/mol or lower.” There’s your relaxed 8.0 per cent target.

And that fits with the findings of ADVANCE and the Veterans Affairs trial. In long—time patients who have already developed atherosclerotic cardiovascular complications, stringent targets may not do much good. Here’s Paul Drury expanding on the theme of personalized therapy.

PAUL DRURY: If you’re talking about an older frail patient, perhaps living alone an older, frail patient—perhaps living alone on an agent that can produce hypoglycaemia—who’s already had a fall or two—who’s had a vascular event–and is struggling to cope with their diabetes—then they should have a higher target because the hazard and risk of serious damage is high. If on the other hand, you have a young patient—age 40, 50 or 60—newly-diagnosed with a long life expectancy, otherwise healthy with no other comorbidities, and you’ve got them on metformin and possibly something else, why ever would you de-intensify? You would absolutely go for the best glycemic control you can achieve without hypoglycaemia. But it really is horses for courses and I don’t remotely agree with the sort of blanket statement in the ACP, which I think it has not been thought through.

SOPHIA ZOUNGAS: So, I think they are just overgeneralising a small risk that applies to a particularly high-risk group of frail people with diabetes where, yeah, it would be more appropriate to be more modest and conservative in the targets. You know the paradigm really now is individualisation. It seems to me that the ACP are well behind the eightball from us where we were already doing this a decade ago.

MIC CAVAZZINI: And even the ACP makes this the first statement of their guideline—that clinicians should personalise goals for glycemic control. So, if everyone’s agreed that personal factors are more important than numbers why the big drama about whether the target is specifically 7 per cent of 7 to 8 per cent?

SOPHIA ZOUNGAS: What you ask is a good question, and you may say we’re arguing semantics about where the numbers are, and should we be using numbers? But unfortunately monitoring glucose levels is all about numbers. People need to understand what a normal glucose range is and what they’re aiming for. If we take that away, then it’s —I use the analogy it’s like taking the tools that a pilot has to fly an aeroplane. If we don’t know what our numbers are, well, you know how are we flying our plane?

MIC CAVAZZINI: What’s a safe altitude to fly at?

SOPHIA ZOUNGAS: Correct. So, you know I think you can’t fly blind and you need to empower those individuals with diabetes and give them the skills the understand what the numbers mean. But also not create anxiety by being overly aggressive with targets or overly aggressively titrating therapy.

MIC CAVAZZINI: Yeah, so personalisation can be tricky in a cohort of patients who tend to have many complications. In a JAMA viewpoint, endocrinologist Professor Kasia Lipska says that each sub-specialist is focused on their own performance targets and might not realise how hard it is for the patient juggling multiple regimens. Trying to meet all of these can be incredibly frustrating for the patient and not actually improve quality of life. And she says that one of her patients told her that recovering from a car crash had been easier than managing diabetes. Is that a familiar refrain among your patients?

SOPHIA ZOUNGAS: So, many do suffer from distress in relation to the amount of care for their diabetes. And as you said, it’s not just about glycaemic targets—it’s about all those other aspects of care: foot care, eye review, blood pressure, you know, “Am I taking my other cardiovascular preventative therapies?” And it’s our role to be able to give them the skills and the confidence to be able to do that, you know, to empower them, and spending a bit of time with the individual understanding what their goals are. If a particular patient’s goal is, you know, “Good quality of life, but I don’t want really to be taking injectables” then you know you’ve got to be able to manage.

PAUL DRURY: It is fair to say that many patients struggle, firstly, to manage their diabetes on a good day, and often it isn’t helped by the impression given by their carers that they’re not trying hard enough. And I think one has to caution good intention there with what is feasible and try and find out what the limitations are in a patient’s adherence and understanding and try and correct those in a positive way rather than setting a gold standard that by definition only a few people are going to be able to meet.

I think we forget that if we say that our target for a patient is 53 many of them will take that literally and 54 is a fail. And I would always recommend that we target a range and that people look at the trends that people’s HbA1c is showing, not look at one result versus the last one. And the other thing that happened in New Zealand was a move away from the international-agreed diagnostic criteria of 48 millimoles, to 50. And the reason for that was that we were seeing people inappropriately diagnosed with diabetes based on a single HbA1c test. Because it’s very easy to get a diagnosis of diabetes but once you’ve got it, it is actually virtually impossible to lose it. And as everybody knows all tests have an element of variability starting with the laboratory variability, which in the best hands is now about two, two and a half per cent, but on a bad day, in the wrong lab, it could be as much as 5 per cent, so you could get misclassified merely on the basis of a laboratory error.

MIC CAVAZZINI: And apart from the variability in the assay itself is it worth mentioning some of the conditions which might confound the blood measure?

PAUL DRURY: Yes. I mean it’s become increasingly clear that there are a number of situations from anaemia to haemoglobinopathies to other circumstances of increased red cell turnover where HbA1c can be in error, and if you’re making a diagnosis in a patient who has no symptoms then you really do need to be a little bit more sure. There are ethnic differences and there are also individual differences, so that an HbA1c of 50 millimoles [per mole] in one person would actually equate to 46 in another person and 54 in yet another person. There is no simple way around that and that’s another reason for wanting certainty that any abnormality that you’ve got is consistent.

One lesson that people sometimes haven’t taken on board is that if you don’t do something to a patient they often stay very much around the same level unless something fairly major happens. And that can be a change in therapy, but equally well if they come across very difficult home circumstances, job circumstances, relationships, what you will, then that can make a major difference, and fiddling with the therapy in that situation is less relevant than trying to help with whatever the social circumstances. And that’s one of the reasons why it’s so important to have continuity of care – that you need to know your patient in order firstly to set a target, and secondly to choose therapies that are manageable and affordable for them.

***

MIC CAVAZZINI: The drugs classically used for glycemic control start with metformin and insulin, which both suppresses glucose secretion from the liver. They act synergistically and don’t typically promote weight gain. If either medication is not tolerated, sulfonylurea drugs can be used to stimulate insulin secretion. All of the above are contraindicated where there is severe renal or hepatic impairment, in which case thiazolidinediones, are another option. These improve insulin sensitivity and protect insulin-producing beta cells, and like metformin don’t, on their own, cause hypoglycaemia.

As we’ve already mentioned, tight glycemic control doesn’t do much to mitigate cardiovascular disease already entrenched after a long disease course. It’s in these patients the new drug classes might have the most benefit. Sodium—glucose transporter inhibitors block renal reabsorption of glucose into the bloodstream. Analogs of the glucagon—like peptide promote insulin secretion, and the enzyme DPP—4 which inactivates GLP can be inhibited to similar effect.

There’s been a flurry of recent trials of these drugs with their own contrived acronyms like LEADER, CANVAS, CREDENCE and SUSTAIN. These have shown that the newer drugs can promote recovery in markers of macrovascular and renal disease in patients with these comorbidities. But, the benefits of treatment in this cohort are measurable after just a few years— much sooner than can be explained by their glucose-lowering effect alone.

When the American College of Physicians revisited glycemic targets earlier this years, the guidelines explicitly did not take into account any of these findings, as the new trials didn’t fit their ‘treat-to-target’ design. But hot off the press, is a consensus statement by the American Diabetes Association and the European Association for the Study of Diabetes.

PAUL DRURY: What’s really new in the recommendations this last month, in that previously there’s been “Here is a list of drugs, but we can’t give you any very definite advice on which you should choose”, now there is a clear recommendation for subgroups. Just looking at the abstract summary that just says, “For patients with clinical cardiovascular disease an SGLT2 inhibitor or a GLP1 receptor agonist with proven cardiovascular benefit is recommended.”

The big caveat I would say is that the benefit that’s been shown for GLP1 and for SGLT2 are applicable to those patients who either have established cardiovascular disease or who have chronic kidney disease, and I think we need to be careful not to assume that that benefit applies substantially to people without those complications. Though I’d actually look around the world and say that the uptake of DPP4 as well as GLP-1 and SGLT2 has been greater than justified by the evidence. If you go to DPP-4 then really there’s not an enormous about of evidence for benefit there and their biggest advantage really is that they don’t produce hypoglycaemia, and as such are a good choice for patients in who hypoglycaemia is a particular no-no for occupational or other reasons, but they’re fairly weak in their glycaemic effect.

MIC CAVAZZINI: These drugs—the SGLT2 inhibitors are also associated with clinically significant reductions in hospitalisation for heart failure as compared to placebo, and they also were associated with improvements in renal measures of glomerular filtration rate and secondary renal endpoints. So, the consensus recommendations are that in patients with these complications those would be a first-line suggestion?

PAUL DRURY: GLP1 receptor agonists are generally recommended as the first injectable medication. But we do have a situation in New Zealand that there is no GLP1 agonist or SGL2 inhibitor currently funded by Pharmac. And one thing that I think is really important, with adherence with any diabetic drug in general the longer-acting your job then the less sensitive to, shall we say, inadequate adherence. But if you have a drug that has to be given twice daily, of any sort, then missing it for six hours becomes really rather important. So, I think there is a welcome emphasis in the new guidelines that actually medication adherence is really important and there’s lots of local evidence that medication adherence for metformin and sulfonylureas and insulin is not good.

MIC CAVAZZINI: And interestingly most of the evidence on these new drugs came from trials of patients who were not meeting their glycaemic targets despite metformin treatment—and there’s no suggestion that these add-on therapies would confer further cardiovascular benefits in patients that are meeting targets. But if you are getting a protective effect with one of these drugs do you still need to worry as much about getting blood sugar down?

PAUL DRURY: That’s a very good question, and I think at the moment you’d have to say, if you’re being purist on evidence, that we don’t actually have the evidence that tells you that, and as a caveat we don’t actually have a lot of evidence on combinations of these various therapies. I think we’re all agreed that the relationship of microvascular events to HbA1c is exponential at least at the top end of the curve and to bring somebody from HbA1c of 105 to 85 actually produces a greater incremental benefit than going from 85 to 64.

MIC CAVAZZINI: Right. In light of that new ADA and EASD statement, which is very thorough and comprehensive, for the ACP to come out in April it seems almost anachronistic to focus so much attention on the glycaemic targets when there’s so much else going on, but in your view this kerfuffle hasn’t really influenced Australian or New Zealand physicians; they’re getting on with...?

PAUL DRURY: I think, firstly, New Zealanders don’t tend to read the American College of Physicians Guidelines, they’d be much more likely to go with the American Diabetes Association and the European Association and indeed the Australian Association. Maybe we need a mechanism to more formally adopt an external guideline. And certainly it’s important that all those who care for diabetes and particularly GPs, specialist nurses and practice nurses, are all singing from that same hymn sheet. Because one of the great criticisms that patient groups give is that they get different messages from different carers.

***

MIC CAVAZZINI: It’s well worth reading that consensus report published in the journal Diabetes Care in September. It describes where all medications, new and old, fit into the full package of lifestyle advice for treatment of Type 2 diabetes. The 2014 blood glucose management algorithm of the Australian Diabetes Society also provides a very rational approach for stepping up combination therapy. Sophia Zoungas offers some final considerations.

SOPHIA ZOUNGAS: The algorithm doesn’t necessarily advocate one class over another—the first approach is add the agent, titrate the agent to its maximum tolerated dose and then add a new agent. What’s I think different about the Australian Diabetes Society treatment algorithm is that it asks the clinician to consider whether the agent that was added was effective in any way and potentially to stop that agent if it wasn’t shown to be effective.

MIC CAVAZZINI: Yeah, how soon would you expect to see an effect before either stepping up or alternating?

SOPHIA ZOUNGAS: It will depend on the agent. The new agents that have an immediate effect of onset, you should certainly expect to see an effect within the first three months. And if there’s no effect you could say the particular individual is a non-responder. you know between 20 and 30 per cent will be non-responders. In the case of the older agents such the glitazones one certainly had to wait at least three months due to their slow onset of action.

Clearly when you add the second and third agent the amount of change in A1c may not be as great as the first agent, and as you get closer to target we know that the proportional reduction is dependent on where you start from, so it’ll be smaller, the total effect of that particular addition. And what we’re seeing with these new classes of drug is that there are potential glucose—independent effects, and then one would suggest that the target is not part of the decision—making process rather the assessment of cardiovascular risk.

MIC CAVAZZINI: Yeah. And a few of the commentators have made the comparison to lipid-lowering strategies in the past where clinicians used to treat to target and now treat to risk. Was that a hard cultural shift to make and will that shift be required in the diabetes management field?

SOPHIA ZOUNGAS: I think the difference between lipids and glucose is there is a lower glucose level that causes harm, and yet for lipid we don’t know about that or we think there isn’t. We’re still developing the evidence that will inform the way these new drugs will apply and particularly whether they specifically to apply to more high risk, where there’s a more absolute immediate need for prevention of further events.

The important aspect to all these things though is that these agents is that these agents do have a glucose lowering effect. So even if we were to apply them without having an HbA1c target, we would still need to measure glucose levels if they were being used in conjunction with other agents that may lower glucose levels as well and potentially cause hypoglycaemia.

So, they’re not a set-and-forget group of drugs, and in fact, you know we do have a lot of people now who are also having other therapies whether it’s for management of mental health issues or other comorbidities where you know use of these agents, you know, steroids, antipsychotics, can acutely increase glucose levels and contribute to significant, acute metabolic decompensation. So, I think it’s a matter of balance, and we don’t want to go so far the other way we completely forget that there are aspects of glucose management that go beyond long-term vascular complications, and need to think about the other therapies that we’re using as well.

***

MIC CAVAZZINI: Thanks to Sophia Zoungas and Paul Drury for guiding us through this tangled story. The views expressed are their own, and may not represent those of the Royal Australasian College of Physicians. It’s a lot to take in, so please go to our website for links to all the references mentioned, particularly the Australian prescribing algorithm sketched out on the Diabetes Society website and summarised in MJA. Our address is racp.edu.au/podcast.

You’ll also find an audio appendix with interesting interview outtakes, and some video lectures in the College Learning Series from two of our top endocrinologists. If you like Pomegranate Health, please tell your colleagues to subscribe via any podcasting app, or the mailing list found on our web page. The podcast can now also be streamed from Spotify.

Finally, over the next couple of months, we’ll be recruiting a new podcast editorial group to start in February. The role of the group is to provide feedback during the development and post production of every episode. It’s fun, it’s all done by e-mail, and it’s worth a few CPD credits. There will be room on the group for a couple of health professionals who are not members of the RACP, and we’ll be seeking perspectives from the College’s Consumer Advisory Group as well. There are expression of interest forms at the website, and you can send them back to podcast@racp.edu.au.