Good morning, everyone. Welcome to this session of Hot Topics in Endocrine Hypertension. I'm Adina Turku. I'm an associate professor at University of Michigan, and my co-chair is Dr. June Yang. She's the head of Endocrine Hypertension Group at Hudson Institute in Australia. So our first speaker today, and by the way, I invite you to scan this code, and then you will have a chance to ask questions in between each talk. So our first speaker is Dr. Reena Bankosh. She's an associate professor at Mayo Clinic, Rochester, and her talk is on maternal and fetal outcomes in pheochromocytoma in pregnancy. Thank you. Okay, someone is opening for me, which is great. Hi, everyone. Welcome to the session, and I'm glad I'm the first one to go because my talk has a very different flavor from the rest of the talks you're going to hear. This is the talk which is based on very little evidence, and it will be more clinical in the hopes that whenever you do see a patient with pheochromocytoma in pregnancy, you'll remember this. These are my disclosures. None of them are related to this talk, and we have three objectives. To describe the characteristics, the clinical course, and the outcomes of women with pheochromocytoma and paraganglioma that I would be calling PPGL during this talk, during pregnancy, to determine the predictors of maternal and fetal outcomes in women with PPGL, and to discuss the best approach to management of PPGL during and hopefully before pregnancy. This group does not need reminder of what PPGL is, but what we're worried about is those episodes of catecholamine excess, which can lead potentially to severe clinical consequences. In general, there is 40 to 50% genetic association with PPGL, but in young women, because we're talking about pregnancy during this talk, obviously in young women who have PPGL, the likelihood of genetic association is probably higher than 40 to 50%, and pregnancy and PPGL is a very rare occasion. Most of my data and my information will originate from this study published last year in Lancet, Maternal and Fetal Outcomes in Pheochromocytoma in Pregnancy. It is an international retrospective multi-center study with 28 countries across the world participating, multiple centers within those countries, and we also decided, because of the rarity of the disease, to add systematic review of studies published between 2005 and 2019. We included only case series of at least five patients, or five pregnancies, I should say, and we included only women with pregnancies between 1980 and 2019, meeting one of the two criteria. PPGL was discovered before or at any time during pregnancy, or PPGL was discovered within 12 months postpartum, and we included the second criteria because we all know that PPGL is a slow-growing disease. Our outcomes of interest included maternal and fetal death or maternal severe cardiovascular complications of ketocholamine excess. And these are the numbers of patients originating from the multi-center collaboration, and that's the second column, and from the systematic review. We had 197 patients from around the world, and 52 more added by several papers in the systematic review for a total of 249 pregnancies. These 249 pregnancies happened in 232 patients. The difference was 10 patients had two pregnancies with unrecognized PPGL, one patient three, and one poor patient had six pregnancies with unrecognized PPGL. So what did we find out? We found that the majority of these patients had one single pheochromocytoma or one single paraganglioma during pregnancy. Some 19 had bilateral pheochromocytoma, and then a minority of patients had multiple PPGL or metastatic PPGL. The median tumor size was 5.3 centimeters. And 95% of our patients in our cohort had functioning PPGL. And looking at the color coding for this graph of function, half of our patients had severe ketocholamine excess, 10 times above upper normal range for the 24-hour urinoplasma metanephrines. Another third or so had moderate ketocholamine excess, anywhere between two and 10 times above upper normal range. Only 62% of patients were tested, partially because it's a study that started in 1980, but also because some patients refused or it was not available. And of those patients who were tested, the majority had either MEN2 or SDHB, followed by VHL and SDHD. So here's our cohort. We had 249 pregnancies with 15% of them actually diagnosed, well, in 15% of them, PPGL was diagnosed before conceiving, before conception. 54% were discovered during pregnancy, and 31% of patients were diagnosed with PPGL only after they delivered. Looking at patients who were discovered with PPGL during pregnancy, the median gestation week of discovery was 24 weeks. And looking at patients who were diagnosed with PPGL only after they delivered, it ranged quite a bit between right at the time of delivery, someone thought, okay, maybe I should suspect ketocholamine excess up to the whole one year after. But median was six weeks. We've excluded patients with non-functioning PPGL from further analysis. So 231 pregnancies were analyzed. We've looked at alpha-blockade, which was used only in those patients with recognized PPGL prior to delivery. 63% of patients were treated with alpha-blockade during pregnancy, and 25% of those patients had surgery to remove PPGL during pregnancy. Overall, we found the rate of complications in babies was 9%, and in mothers, 8%. But when we looked at the rate of complications based on the timeline of discovery, we found that 0% of complications were noted if PPGL was discovered prior to conception. 8% of complications in babies, and 3% in mothers if PPGL was discovered during pregnancy. And the majority of complications were found only in misperigangliomas during pregnancy. 13% in babies, 19% in mothers, for a combined overall complication rate of 14%. So before doing the analysis, we actually predefined the factors we thought would be good to look at, and that was timeline of diagnosis before or during pregnancy, degree of catecholamine excess, location of PPGL, whether the patient has metastatic or non-metastatic PPGL, presence of genetic predisposition, and therapy. We looked at alpha-adrenergic blockade and surgery. We also looked at type of delivery, concentrating on two main ones, C-section and vaginal. Okay, so let's look at those patients who were diagnosed before pregnancy. So these are the patients who knew they had PPGL prior to conception. So 17 patients had metastatic PPGL, 20 patients were diagnosed with PPGL, which was mainly a single-fuel chromocytoma, but did not yet have surgery. I would say as a group, they had milder catecholamine excess, and that probably altogether allowed appropriate planning for therapeutic procedures before pregnancy in some patients where it was possible, early initiation of alpha-adrenergic blockade in those patients, and I would say as a group, these patients had probably higher number of appointments with their physicians and higher level of expertise and monitoring by their team. And I should say most of them planned to be pregnant. It was not a surprise. Looking at patients who got diagnosed with PPGL during pregnancy, when we compared the complication rate of these patients to those who were diagnosed before pregnancy, it was still much higher risk, 8.3 times higher risk of complications. The reason for that was absent or inadequate alpha blockade. For example, one patient was treated only with a small dose of alpha blocker and only for one week. Complications during combined C-section and adrenalectomy for one patient, and then late diagnosis for most patients for when complications already developed. And finally, looking at the most severe group, those who was missed PPGL during pregnancy, PPGL was not recognized in those patients until after delivery or actually during delivery. No action was taken for this, obviously, and those patients usually had some sort of emergent procedures. Obviously, none of these patients were on alpha blockade because it's not the usual treatment of hypertension. And as this systematic review showed, frequently, these patients are misdiagnosed as preeclampsia, and this systematic review by Langton et al, 20% were misdiagnosed as preeclampsia. So this, the complicated graph that I will a bit digest with later slides, looking at panel A, these are all patients diagnosed before, during, and after pregnancy. And then the panel B below looks only at patients with recognized PPGL prior to delivery. We decided on this separate analysis because we wanted to understand whether the risk changes if a physician and patient actually knows they have PPGL during pregnancy. So not surprisingly, if the PPGL was missed during pregnancy, we have a much higher risk, for example, 27 odds ratio for those diagnosed after pregnancy, and 8.3 during pregnancy when compared to those with diagnoses made before pregnancy. But what I think is more interesting is, if you look at panel B, you'll see that certain factors we may expect to be associated with complications were not. For example, the timeline, earlier in the study versus later in the study, did not have much difference. C-section versus vaginal delivery did not have any impact on complications. What did have an impact is the degree of catecholamine excess, and the two protective factors were presence of genetic association, presence of metastasis, which could be counterintuitive, but those are the people that were more likely to be recognized earlier with some sort of therapy instituted earlier. Finally, surgery did not have any impact on complications if done during surgery, during pregnancy, and alpha adrenergic blockade had a significant protective impact. A little bit more on C-section versus vaginal. So 70% of patients had C-section, and only 30% of patients had vaginal delivery. And these are the patients where there was a choice. People did know they had PPGL. The medical team did know that the person had PPGL, and so C-section was more commonly chosen in those patients. When we looked at why, degree of catecholamine excess was higher in patients undergoing C-section versus vaginal delivery, and those treated with C-section were more likely to have an abdominal or pelvic PPGL, possibly because physicians were concerned about the contractions and gravid uterus pressing on PPGL in the same location. But even those with vaginal delivery had slightly more similar location in abdomen and pelvic. So these groups had same maternal age, same tumor size of PPGL, and same rate of metastatic disease, and they also had same rate of complications. This data should be interpreted carefully because clearly there were some other factors individualized to the patients when C-section versus vaginal delivery was chosen. But one take-home point is that vaginal delivery can be safe in carefully selected patients. We also were interested to understand why mechanical factor is actually not the main reason for those catecholamine surges and gestation loss. We found that there is similar percentage of fetal loss at earlier gestation age versus later gestation age. And also another systematic review and meta-analysis confirmed that. So we thought maybe there's some other contributing factors to the pregnancy loss early in gestation other than mechanical factors. And recently there was a study published by Lopez who found presence of LHCG in pheochromocytoma. They did a nice little study where they showed that LHCG stimulated epinephrine secretion from pheochromocytoma cells, possibly explaining the catecholamine surge and crisis early in the pregnancy leading to gestation loss. So I wanted to summarize some of these factors from our study and a few other smaller studies. Timeline for diagnosis of PPGL, especially when it's done before conception, is probably the most important thing we can do. Degree of catecholamine excess does matter with a higher degree of catecholamine excess impacting or influencing a higher rate of complications. Location of PPGL matters. Abdominal or pelvic location is much worse than PPGL elsewhere. Metastatic PPGL in our study did demonstrate a lower rate of complications. But remember multivariable analysis was not possible because there were not so many outcomes in the study. So these are the patients also with well-treated disease and with smaller tumor burden and with lower catecholamine excess as well. Presence of genetic predisposition also was associated with lower complication. Again, possibly because those patients were carefully selected to become pregnant. And then therapy during pregnancy. Surgery had no effect on complications and the placebo alpha-adrenergic blockade definitely was protective if it was optimal as far as dose and duration. Finally, type of delivery. C-section versus vaginal had no impact. This decision needs to be individualized based on all of the above factors and every patient is different and most of us would not be seeing hundreds of these patients so multidisciplinary team involvement is important. So to conclude this section, recognizing PPGL before delivery is key. Patients with metastatic PPGL can have a safe pregnancy if under close monitoring. Surgery during pregnancy can be safe but may not be necessary, especially in patients with extra abdominal PPGL. So this should be individualized to specific situations and only performed under optimal alpha blockade. Alpha blockade prevents adverse outcomes especially if optimal as far as duration and dose and choice of C-section versus vaginal delivery is individualized. I want to also just provide several more slides on recommendations for preconception planning, women for risk of PPGL. This is absolutely based on expert opinion because there are no studies randomizing women for different treatments. But I do want to point out that advising or recognizing patients at risk of developing PPGL in the reproductive years is the most important thing. As such, in women with known germline variant associated with PPGL, biochemical testing early or before decision to become pregnant and even cross-sectional imaging before conception should be considered. In a woman with personal history of PPGL, biochemical testing, cross-sectional imaging and also genetic testing if it was not yet done should be considered. And in a woman who has family history of PPGL, biochemical testing and genetic testing for germline variants could be considered. In those who did become pregnant and have PPGL, alpha adrenergic blockade, beta blockade and calcium channel blockade would probably be the most effective medications though not approved for pregnancy, probably safe during pregnancy. And this is it, thank you. I just want to add one more thing that this was a huge study, it was 114 co-authors who worked very hard to contribute their data. So my biggest acknowledgement goes to the International Multicenter Pregnancy Workgroup. Thank you. Thank you, Dr. Bancourt, this was a really nice talk, and Dr. Lacroix, why don't you take the first question? André Lacroix, Montreal. Congratulations for conducting that study. That's very important hard work. Since you cited the data from Elie Wieler-Fein's group on the effects of LH receptor, they conducted a study also that showed that this was true mostly in patients with cluster 2 PPGFOs, where 40% of the patients expressed the LH receptors. So retrospectively, can you see within your cohort whether patients who had cluster 2 mutations, was there more deterioration of the outcome during pregnancy in those patients compared to other subtypes of genetic mutations? Oh, this is an excellent suggestion. I have not done that, but I absolutely will. Hi, Tim Koolvaar from Rossmans University, Rotterdam. Thanks for your brilliant work and presentation. So I was a bit triggered when you said that pheochromocytoma was often misdiagnosed as preeclampsia, because I think there's a very fine line between having superimposed preeclampsia because of the hypertension associated with the catecholamines versus it being a misdiagnosis, and you have, you know, because of the catecholamine excess in itself signs that could be diagnosed as preeclampsia. And I was wondering if you also have any data on aspirin treatment, because of course aspirin is very well capable of preventing preeclampsia. So I personally think that every woman who becomes pregnant with a secondary cause of hypertension or underlying disease that is a secondary cause of hypertension should have aspirin prophylaxis. I'm just wondering if you have any data. Yeah, so I mean, all good thoughts and questions. We were not able, so first a little bit about our study and what was available. The data on medications and full treatment of this woman was not available. In fact, we've initially tried to consider all medications for hypertension at least to understand whether alpha block it by itself or other medications can be looked in a subgroup analysis, and that was not possible because of missing data. And that extends to other medications such as aspirin. So from this data, I'm not able to conclude that. Now, the second part, the second comment to your question is how do we know whether it was ketocolamine crisis or preeclampsia or both? We don't. We just know that a lot of those women presented with significant episodes of hypertension that sometimes were diagnosed as preeclampsia only, because PPGL was not diagnosed. So what we need actually to answer the questions you are raising, which are very important, is a prospective study. And I don't know how feasible it is. I hope it is. I'm sure you'll find a way. Thanks. Dr. Bankos, just a couple of questions from online. Is the higher proportion of functioning PPGLs during pregnancy, difficult is diagnosing non-functioning PPGLs, or is there something about pregnancy that activates the PPGLs? Well, I guess there's at least one study showing that silent pheochromocytoma can be activated by pregnancy, but I would say it's just most of the women don't know they have non-functioning PPGL during pregnancy. So we are probably, we have a selection bias towards functioning. And another question is, are there any conclusions from your study about the safety of different types of alpha blockade or the addition of beta blockers and calcium channel blockers? Unfortunately not. We tried our best, but no, we were not able to do the doxazosin versus phenoxybenzamine analysis. I guess it's your rare condition, so it's hard. Yes. Quick connected question to this. So in women who are diagnosed with pheopera before pregnancy and were treated during pregnancy, you show no complications for either the mom or the baby. So no complications for the baby despite treatment. But for those who were treated or diagnosed during pregnancy, there was a higher risk of complications for babies than for moms. So I was curious, what were the complications for the babies? And were there any deaths in either babies or moms? Yeah, so all fetal complications were death. No babies survived. So when you see complications for babies, they're all death. For moms, the majority of complications were non-fatal complications. And some of them were pretty severe, with strokes and multiple transfusions. But if I remember correctly, three mothers died. But majority of maternal complications were not. Now, I did want to add that with treatment, when a patient, one of the patient at least, I should say one of the baby died during surgery while attempting during pregnancy. So not all treatment is completely harmless. Okay, we're just on time. Thank you so much again for your talk. Thank you. Thank you. Next, we invite Professor Morris Brown, who's a professor of endocrine hypertension at St. Bartholomew's Hospital Medical School in London, the oldest hospital in the UK. He moved there from Caius College, Cambridge, following a path taken by William Harvey about over 400 years ago. So thank you. Well, thank you very much, June. And it's wonderful to be back in front of a live audience. I'm going to share with you the excitement of this story of how somatic genotype in artisterium-reducing adenomas can influence the presentation and outcome in a subset of primary artisterianism. And I'll talk briefly about the impact of those discoveries on other patients. I have one disclosure, which does not bear on what I'm going to tell you. So like many stories about somatic mutation, mine in a way goes back to the landmark paper of Rick Lifton in 2011, when he first reported the somatic mutations in KCNJ5. And it was John Funder's editorial in that science paper, which in part first prompted our thoughts about genotype-phenotype differences. And John remarked on maybe the atypical large size of the tumors which Rick had studied, that may not be correct, but when, like many groups around the world, we went to our freezer and took out some 50 or 60 of the adenomas and sequenced them, we plotted the data in this, what I came to call, boxing ring diagram, plotting the size of the tumor on the y-axis against the age of the patients on the x-axis. And you can see in the red corner, top left, that the patients with the KCNJ5 mutation are mainly younger women, whereas in the blue corner, bottom right, the patients are mainly, not sure if the laser's working, the patients are mainly men and presenting with rather smaller tumors. So when we look more closely at the tumors, we could see that the tumors in the top left, the KCNJ5 mutants, paradoxically the cells resemble the cells of the zona fasciculata, rather large, remember the high level of the enzyme 17A1, whereas the cells of the tumors in the bottom right-hand corner look more like the normal compact cells of the zona glomerulosa. So we took 10 of the latter tumors and sent them for whole exome sequencing. And this table, which was mainly drawn to show the characteristics in the left column of the patients, mainly males, and the next column the size of a tumor, all less than two centimeters, some less than one centimeter. Nine out of 10 of those tumors had a somatic mutation, again a function mutation, in two very, very well-known ion transporters or ion channels. But the subject of interest today is the odd one out, the patient at the bottom, one of the only two women in the cohort who had a mutation in beta-catenin. And shortly after we published the paper, a second patient presented whose sequencing, again, drew our attention to a serine mutation in exome three of beta-catenin. And the phosphorylation of this residue determines whether or not the beta-catenin gets to the nucleus and directs differentiation of the cells. And although these patients presented almost 10 years apart, the common theme between them was that both had presented in pregnancy. And our good fortune was that the index case, we had used her tumor in a microarray study of 20 tumors comparing expression of every gene in the genome between tumor and adjacent adrenal. And all we had to do was resolve the spreadsheet calculating the difference between transcript level in the index case and the other 19 patients. And the gene which was most differently expressed was the gene encoding the receptor for luteinizing hormone or HCG pregnancy hormone. So we confirmed this on qPCR. We're now looking at a log scale both in the index case on the left, the second patient in the middle, and a third patient whose sequencing we'd now looked at, a patient who had presented in early menopause. And the manifold increase in LHCGR in these patients contrasts with 13 controls with KCNJ5 or other mutations. So Ada Teo who was MB PhD student with me at the time wrote this up as a brief report in the New England Journal. But this proved to be at most end of the beginning rather than end of the story. And when I moved shortly afterwards from Cambridge to St. Bartholomew's in London, Queen Mary University London, I met this 12 year old boy who had presented to cardiologists with chest pain and been noted to have severe hypertension, low potassium, high aldosterone, and low renin. And that drew attention on a CT scan to a right adrenal adenoma, which was removed and cured all the boy's problems. And some years later, on the right is the tumor stained with Celso's wonderful CYP11B2 antibody. The tumor was sent for whole exome sequencing and on so-called MUTEC2 software analysis of the DNA sequencing, compare the tumor with adjacent adrenal or germline. We sent this spreadsheet of all of the possible mutations and standing out was butyrgitinin. And I realized the boy had probably presented in puberty and indeed the pediatrician later confirmed this. But running my eye down the list, I saw at the bottom GN11, which is the G protein or one of the two G proteins in the adrenal, which mediates the aldosterone response to angiotensin and was an obvious further candidate for the hyperaldosteronism. And the actual residue which is mutated, bottom right hand corner, Q209, when I looked at the literature, I realized it was a well-known hotspot, both in uveal melanomas, where it's mutated in over half the patients, and in certain congenital skin lesions, such as those found in Sturge-Faber syndrome. So we went again to our freezer, sequenced as many of our tumors as we could, and most of them had no mutation in GN11, but all seven tumors by this time we now had with mutations in exon three butyrgitinin, also on the bottom row had mutation of the Q209 residue, either to histidine or to proline. And when we transfected H295R cells with either of these mutations, the 295R cell already has the exon three butyrgitinin mutation, a combination led to 10, 20-fold increase in expression of CYP11B2 and secretion of aldosterone. But the bit of the phenotype, of course, we were most interested in, was what the W mutation did to LHCGR expression. And this is plotted on a log scale, and you can see that the seven patients with the W mutation of, sorry, yeah. The seven patients with the W mutation of butyrgitinin and GN11 in the tumor indeed had a very large increase in LHCGR compared to controls either adjacent to adrenal or the other tumors with different mutations. Now, I keep saying GN11 or GNAQ are very similar G proteins, and you'll see as the story progresses that we saw tumors with both types of mutation. But because, remarkably, all seven of our patients had somatic mutations of both genes, we were not able to answer the question whether we needed both to see the increase in LHCGR. So I knew that Christina in Paris had sequenced a large number of tumors and had several with butyrgitinin mutations, and so she joined us, and Fabio sequenced all their butyrgitinin mutations and many of the others to show that, indeed, five of theirs had a W mutation either of GN11 or GNAQ. And when he measured the LHCGR expression, found it was only those with a W mutation in red that had high expression of LHCGR, and none of those with solitary mutation of butyrgitinin. And this wasn't just true at RNA level. It shows that Booker in Paris did the immunohistochemistry, and this slightly atypical tumor where there was patchy expression of CYP11B2, you can see that where the CYP11B2 is high in the bottom panel, there's also high expression of LHCGR. In my lab back in London, Alina and Sam did the transfection studies using primary zoonic megalozoa cells, and the panel on the left is three different looks on an overlay of the same field of cultured cells looking into different immunofluorescence channels so that you can see the expression of transfected GN11 in green, CYP11B1 in purple, and then the question whether the transfection leads to expression of LHCGR, and I've blown up the top right-hand panel so that you can see it's only a cell which is both green and purple in the top right of the field, but it also expresses LHCGR. And this was repeated multiple times so we could quantify this increase, and on a log scale again, you can see on the far right column it's only cells transfected with both the mutant beta-catenin and GN11, which causes the manifold increase in LHCGR expression. So we're pretty confident in saying that the phenotype of presentation in pregnancy, puberty, and maybe early menopause is because the LH is acting on a very high level of receptor. This receptor is coupled to adenylate cyclase, and cyclic AMP is one of the known stimuli to aldosterone production. Well, LHCGR is actually not the only gene which is upregulated in these double mutant tumors. This is a heat map from a pooled set of microRNA-seq studies which included between them three patients who had the double mutant. Not sure why I can't get the point. So the double mutant is the one which is standing out as the red towards the right of either the left hand or the blown up version of the heat map. And you can see that there's multiple, for those who are not familiar with the heat map, the samples are clustered along the top by unsupervised cluster analysis. So if the genotypes line up, that tells us that the transcriptions differ by genotype. And on the right-hand side are the different genes, again, unsupervised clustering. And you can see multiple genes on the right-hand side where I've put arrows against ones of interest that are upregulated in red in the double mutant tumors. LHCGR has the red arrow. But one which may be even more interesting is a gene called TMEM132E, which is a deafness gene encoding a single-pass membrane protein that may be involved in trafficking of G-protein-coupled receptors to the cell membrane. And Fabio, again, looked by qPCR at the level of TMEM expression in the double and solitary beta-catenin mutated tumors in the Paris set. And you can see this, again, enormous increase in TMEM expression only in the double mutant. And in my group, Julia transfected 295R cells to show that TMEM is the most upregulated gene when both mutations are present in the cell. So there's a final bit of laboratory work from Ms. Doy I want to tell you about, which is that some of you as an endocrine audience may have twigged that the Q209 glutamine residue in GNA11 or GNAQ is analogous to the Q227 glutamine, which is one of the two which is mosaically mutated in McEwen-Albright syndrome. And in that case, on the bottom, I show that the gain of function of cortisol production in the adrenal of these patients is due to ACTH acting through GNS and producing more cyclic AMP, whereas for GNA11, it's the equivalent glutamine leading to constitutive production of aldosterone in response to angiotensin. And by a nice coincidence, when I looked at the literature whether there was any evidence of mosaicism for GNA11, I found this paper of whom the senior author, Veronica Kinsler, was one of my first students in my Cambridge days, and she encouraged me to go looking for mosaicism in the adrenal because she had discovered that people with various skin conditions have mosaicism of GNA11 or GNAQ. So we went back to the index case, whose adrenal you saw briefly earlier on, and to a couple of the others. And in the four right-hand panels, you can see different looks at a bit of the cortex near the zone of glomerulosa, illustrating the quite marked hyperplasia that we see in all these patients, but with no CYP11B2 expression. Interestingly, one of the things they do express there is LHCGR, that bottom left-hand panel on the top. And I was interested in what Irina said about LHCGR in the Pheos because the other place in the adrenal we see LHCGR is the adrenal medulla. And then on the bottom, we use laser capture micro-dissection to take cells for extraction of DNA, RNA, for multiple sites around the zone of glomerulosa. And when we sequence these, we can see that in blood, the patient is wild-type for both genes. In the APA, there's a heterozygous mutation of both genes. But if we look at two discrete areas of the zone of glomerulosa, the beta-catenin looks wild-type, but there's either heterozygous or homozygous mutation of the glutamine residue. This was repeated and confirmed in three of our patients using NGS, or Fabio used DD-BCR, in Paris. So we're confident that there is this multiple adjacent mutations. What we haven't absolutely proven is whether it's truly mosaicism, meaning it's present also in another tissue than the adrenal cortex. So we wrote all this up in HGenetics last year as an opportunity to thank many of the people who took part, and there's more of a story I don't have time to share with you today. But there's one aspect I do want to share and finish with, which is the outcomes from adrenalectomy in these patients. So in this table, you can see that by the time of publication, we had 10 patients, all of whom had the beta-catenin and either GN11 or GNAQ mutation. But of most interest is that when we look at the blood pressure after surgery, all 10 had come off all treatment, and the blood pressures, which, not just normal, but super normal. And so our postulate would be either this reflects the short degree of exposure to arthrosteroids, so we don't have time to develop secondary vascular changes, or more interestingly, that because of the mosaicism, we can be confident that the mutation has arisen just in one adrenal and that the other is completely normal. And the question is whether this is telling us anything that can be extrapolated to other patients with primarist duodenism. So I want to finish with a couple of slides from a clinical study that we finished and is hopefully going to be coming out shortly, where we validated the use of our PET-CT scan against adrenal vein sampling, and that gave us an opportunity to look at the genotype of the 78 patients who came to surgery and relate those to the clinical outcome. So you can see here that for the first time, we have more patients with non-Casein J5 than Casein J5 mutations, partly possibly because of the use of two techniques for lateralization, and partly because the red part of column one shows, as Bill Rainey reported a couple of years ago, the black patients, who are about 30% of our study, are much more likely to have catkinase than Casein J5 mutations. But the main question of interest is how these relate to outcome, and here I've re-plotted the data, only now showing those patients who came off all of their antihypertensive treatment with average home readings less than 135, 85. And you can see that among the common genotypes, effectively, only Casein J5 is left standing, with 14 out of the 18 completely cured. And there were two patients in the study who turned out to have double mutation of GNAQ and Casein MB1, who also completely cured. And on a logistic regression analysis, it was genotype which trumped everything else, including age and gender, as predictors of complete cure. So what's particularly exciting is that with the help of Veep Karaut and her team in Birmingham, we collected 24 year old at the start of the study in all of the subjects, and she was able to show on the y-axis the ratio of one of the hybrid steroids to cortisol. There's almost clear blue water between the high levels in the Casein J5 mutant patients, due to the presence of both CYP11B1 and B2 in the same cells, and patients with all of the other mutations. And the literature on this has been a bit controversial, but we feel that this should be replicated and could be used now to predict in advance patients likely to be cured post-adrenalectomy. Well, that doesn't mean others shouldn't have an intervention, and prompted by the debates on ablation versus surgery yesterday, and an old advert, which some of you may remember, I'm tempted to say that if you've got primary aldosteronism, having Casein J5 genotype may be priceless, but for everyone else, maybe radiofrequency ablation is the way to go. But it needs a clinical trial, of course, and so I unashamedly put up a bit of promotion for the trial we're about to start, and anyone who has patients with access to London, please get in touch with us. They've got to have unilateral aldosteronism, hyperaldosteronism, and a proven adenoma. So, in summary, I've shown you that there's a striking subset of primary aldosteronism in which somatic mutation of both the G protein, which mediates the aldosterone response to angiotensin and beta-catenin, leads to very high levels of LHCGR in the tumors, that we think these tumors have probably been lying dormant until they see the agonist at times of high LH or HCG. And the second part of the picture, which we think is of interest, that all these patients were completely cured of their hypertension for one or other of the two reasons that I mentioned earlier on. The slide shows at the top the wonderful group of PhD students who did almost all of the laboratory work, and the degree of divine guidance that we are lucky to share, whether it's St. Paul's Cathedral on the left, or the Temple to Poseidon on the right, when the group came out after a great day at the International Society of Hypertension in Athens. But there were many other people, especially in the clinical studies, who participated in the studies, and who will help with the clinical study, which I've just mentioned to you. Thank you. Thank you, Professor Brown, for a fascinating talk. We only have a minute and a half for a few questions. So shall we start with, yep, Professor Lacroix. André Lacroix, Montreal. This is marvelous work. As you know, we have studied patients systematically looking for aberrant hormone receptors in primary ALDO, and we do find that 30, 35% of patients do respond to LHRH or to LH itself. And we did not find any beta-catenin mutation in the patients we had characterized in vivo, but not looked for GNA11 co-mutations. So I have two questions. Within the genes which are, because some express LH receptors and others GNRH receptor. So in your study, did I see that GNRH was also overexpressed in some of those tissues? And secondly, in those patients, do they have unique adenomas? If you look besides it, are there, are there micronodules with other mutations which are occurring within the same adrenal? Okay, well thank you very much for those questions. So many of them also have high levels of GNRHR, but in our transcriptome data, that's much less specific in that we find that also in many of the other non-KCNJ5 mutant tumors who don't present at times of high age. So my inclination is to think that's less relevant, but I showed you the heat map to say, I don't think this story's finished. In answer to your second question, we specifically don't see ABCCs or other tumors. The ZG is very quiet, and you saw just one of the adrenals. But CYP11B2 is conspicuous by its absence from the rest of the adrenal. But it might be that other patients that also express LH receptors, it's just not the same mechanisms. It could be different. It doesn't seem to be exclusive. Yes, and it's interesting that AHCGR was present in mesonic myeloma of these patients, and on the immuno which Shiraz did for us in Paris. I wouldn't say we're completely sure whether this is a feature of primarily sterilism ZG, or whether all adrenals in patients normal or fears, et cetera, also have AHCGR in the ZG. Something I'm interested in. One very quick question. Thank you. Clifton Bly, Sydney. It's a beautiful story. It'd be nice to close the loop. Have you had the opportunity to use GnRH agonists to show the resolution of hypertension in these patients? It's an interesting question. You know we haven't done that. Thank you. Thank you again. So our next speaker is Dr. Anand Vaidya. He's an associate professor and director of the Center of Adrenal Health Disorders at the Brigham and Women in Boston. He will discuss the unrecognized prevalence of primary aldosterism. Are you doing it? All right, hello everybody. Good morning. It's a pleasure to be here in person. Thank you for the invitation to speak with such a brilliant panel of speakers. As has just been introduced, I'll be talking about primary aldosterism and its prevalence. Here are my disclosures. And I'm not sure if I'm supposed to, not sure if I'm supposed to show this or not. Thank you. Okay, so before talking about prevalence, let me start with a more basic question which is what is primary aldosterism? And this may seem like a trivial question, but I think it's fundamental to answer this question if you're going to assess prevalence. And I bet all the experts in this room have a slightly different definition of primary aldosterism, so let me start by giving you mine. Rather than thinking about primary aldosterism as a discrete disease, I think about it as a pathophysiologic syndrome, a syndrome of inappropriate, that is non-physiologic, relatively non-suppressible renin and angiotensin II independent aldosterone production. And then this aldosterism can cause excessive activation of the mineralocorticoid receptor in the kidney, inducing a vicious cycle of volume expansion that can then consequently increase blood pressure, calories, urinary acid excretion, and it can induce cardiovascular and kidney disease in part due to hypertension, but also above and beyond the effects of blood pressure alone. So practically speaking, when you look for primary aldosterism, you're looking for a relative suppression of renin and in the context of this renin suppression, some degree of inappropriate or dysregulated or non-suppressible aldosterone production. But how you operationalize or implement what's written here in red will ultimately determine how you make diagnoses and how you estimate prevalence, and that's where the variability lies. So a quick word about recognition or underrecognition and its relevance to public health. So I think for 20, 25 years, society guidelines, expert opinions, consensus opinions have recommended that all high-prevalence groups be screened for primary aldosterism. For example, it's recommended that all patients with resistant hypertension or hypertension with hypokalemia be screened. So these groups with high prevalence of primary aldosterism should be screened 100% of the time, at least once in their lifetime. But in reality, there we go. In reality, this number is closer to one to 2%, okay? So it's more close to zero than it is to 100%. And if you think about how many cases are diagnosed off these low screens, it's probably a fraction of 1%. So primary aldosterism is severely unrecognized. I think you've heard this message before in this conference. And the main reason for that is we do not look for it. Nobody's looking for primary aldosterism. This is a fairly consistent statistic from all parts of the world. But you might say, well, sometimes I look for primary aldosterism, and that's true. But when we look for primary aldosterism, we use a variety of variable approaches to interpret the results, okay? So what do I mean by that? So when you talk about estimating prevalence, you have to think about the construct that's used to estimate it. Are you considering primary aldosterism to be a discrete disease that obeys a categorical construct or a continuous syndrome? And I'll try to highlight both these approaches and constructs, trying to highlight the advantages and disadvantages of both. So one important point or caveat is that there is no widespread diagnostic gold standard. And this is probably one of the largest causes of variability across the world. For example, as a diagnostician, you might measure the aldosterone to renin ratio. But the moment you obtain this metric, you will mentally superimpose some relatively arbitrary threshold to designate this metric as either positive or negative or high or low. Or, and or, you may look at the aldosterone level itself, but the minute you obtain that value, you will superimpose some arbitrary threshold to consider it high or low or inclusive or exclusive of primary aldosterism. Alternatively, you might use a variety of aldosterone suppression tests, but there are multiple protocols for these, and each one has its own relatively arbitrary threshold to designate it as high or low. Okay, so this is the categorical construct to designate primary aldosterism. It relies on somewhat or relatively arbitrary thresholds, but it's pragmatic, right? Clinicians often need a tool to help guide them, and this gives them that tool. So there are many studies that have implemented this categorical construct to estimate or predict prevalence. I will not discuss all of them, only a few to illustrate this concept. So here's a study from the Italian group, Militaro-Monticoni's group, where they screened patients with hypertension for primary aldosterism using a two-step categorical construct. So all of these patients underwent measurement of aldosterone to renin ratio. To be considered a positive screen, that ratio had to be greater than 30, and the aldosterone level had to be greater than 10 nanograms per deciliter. I would call this a somewhat conservative interpretation of screening. Based on this construct, a certain percentage of them had a positive screen, and if you had a positive screen, you went on to undergo an aldosterone suppression test, and based on relatively categorical cutoffs, these aldosterone suppression tests were interpreted to be either positive, and in diagnosing primary aldosteronism, or negative, excluding the possibility. It's not critical to know what that cutoff was, but that they used a hardline cutoff. So in this case, that prevalence was 6%, and most of these studies have demonstrated that beyond this absolute prevalence, there's a distribution of prevalence. So what you can see here on the right side of this graph is that if you have severe hypertension, that prevalence is much higher, 10 to 12%. That's quite high. But even if you have mild stage one hypertension, these are individuals on the left side who would never have been screened for primary aldosteronism outside a systematic research study like this, the prevalence of categorically overt primary aldosteronism is still 4%, which is quite high. And so what the main takeaway here is not the specific prevalence estimates, but that you can modify these prevalence estimates simply by changing the thresholds that you use. So if instead of using somewhat conservative thresholds, you liberalize these thresholds, you'll get a different answer. Here's a study we did in the United States where we also tested a group cohort of hypertensive patients, but here the approach was to not miss any cases, to capture as many cases as possible, even at the expense of false positives. So here you liberalize the screening criteria, you lower the aldosterone to renin ratio that you consider to be a positive screen, you lower the aldosterone concentration to be considered a positive screen, and the number who are positive will now increase, and then you can implement a conventional aldosterone suppression test with categorical thresholds, and you will get a much higher yield. Here the prevalence of primary aldosteronism categorically defined was 20%, so it got inflated. And as just one more illustration, here's a study by actually Dr. June Yang in Australia, if I can get this clicker to work, where they took a community-based approach. They instructed primary care doctors to screen every new diagnosis of hypertension. So if a patient was diagnosed with hypertension before treating them, they were instructed to screen them for primary aldosteronism with what I would regard as a fairly robust or rigorous technique, two consecutive aldosterone to renin ratios that had to be elevated. If that was the case, then you underwent a standard saline suppression test, categorical black or white. You have it or you don't, and the number's pretty high. 14% had categorically defined primary aldosteronism at minimum because these numbers could be even higher. Another 4% had two positive aldosterone to renin ratios but did not undergo the saline suppression test. They probably had primary aldosteronism. And another 3% had one elevated aldosterone to renin ratio but the second wasn't elevated, so they were excluded as having primary aldosteronism. We can debate what that means, but at minimum 14%, possibly as high as 20%. These are very high numbers. And then finally I'll show you, there we go, that you can even find categorical primary aldosteronism in normotensive individuals. Here are two studies, one from our group in the US, one from the PADD's group in Greece that systematically conducted aldosterone suppression tests in individuals with normal blood pressure, either oral sodium testing or flugocortisone dexamethasone suppression. And when you do this, about 13 to 14% of individuals with normotensive status have non-suppressible aldosterone production that meets the categorical definition of primary aldosteronism. And I'll talk more about normotensive primary aldosteronism in a minute, but what I will say is if you had conducted screening testing for these individuals with, for example, an aldosterone to renin ratio, you would have missed the vast majority of these because they had aldosterone to renin ratios that were not elevated above the conventional thresholds that we typically use. So this is a summary of some studies that have shown how the categorical construct to defining primary aldosteronism is implemented, which is to say that the reported prevalence estimates that come from this approach can be modified. They're variable. They vary based on the population you test, the variable and relatively arbitrary diagnostic thresholds you use, whether you use dynamic suppression testing, which type, what thresholds you use, et cetera. But regardless of these approaches, what we can agree on, I hope, is that the prevalence that are reported are high. They're higher than what was previously considered and for a public health impact, they're all high. So what I'll shift to now is what our group has been doing for the last many years, which is trying to put these relatively arbitrary round numbers that have been used to define primary aldosteronism aside and asking a more fundamental question, which is how common is the syndrome of inappropriate, non-suppressible renin and angiotensin II independent aldosterone production, which is to say how common is primary aldosteronism pathophysiology? And the way we try to interrogate this is by trying to turn aldosterone production off and then quantifying the spectrum of primary aldosteronism pathophysiology that's seen agnostic of conventional thresholds. So in another way of putting this is we're using qualitative or descriptive methods rather than quantifying it based on some arbitrary number. So this is what I would call the syndromic or continuum construct. It's less biased than the categorical construct because you're not superimposing a random number, but it's also less pragmatic, okay? It's more qualitative. So this is a study we conducted in the United States using participants from multiple sites who had all conducted some variation of an oral sodium suppression test. So all of these participants were studied on a high sodium balance, such that their extracellular volume was maximally expanded, renin and angiotensin II were suppressed, and in the context of this suppression, you would expect aldosterone production to be at its nadir. Any aldosterone production that is continued is some shade of primary aldosteronism pathophysiology. And the participants spanned the entire spectrum of blood pressure. We had patients with resistant hypertension, stage one and stage two hypertension, even those who had normal blood pressure. So these are the main findings. All of these participants have suppressed renin, suppressed angiotensin II are in a high sodium balance, and on the y-axis, you see the main metric of the oral sodium suppression test, which is the aldosterone excretion rate over a 24-hour period as measured in the urine. Here are the data in the normotensive people. So each blue line, vertical line, is an individual participant, and they're ordered from lowest to highest. And what I hope you can appreciate, I don't think I can point, is that the majority of these individuals suppressed aldosterone production appropriately, and then as you move to the right, some suppressed less and less, and some absolutely could not turn aldosterone production off. If I had to describe this qualitatively, what you're looking at is a severity spectrum of non-suppressible renin-independent aldosterone production, also shades of primary aldosteronism pathophysiology. And you can see this pattern not only in normotension, but in stage one, stage two, and resistant hypertension. The difference being that the severity of the primary aldosteronism pathophysiology phenotype parallels the severity or magnitude of the blood pressure phenotype. You could look at the same data in a way maybe that you're more familiar with. These are probability density plots where you can see the full distribution of autonomous aldosterone production for each blood pressure phenotype, where as the blood pressure magnitude goes up, these curves are shifting to the right with a right-tailed skew. I personally prefer looking at the data like this, because this is raw, unadulterated, unadjusted data. And I think it highlights two things pretty nicely. First, it shows that primary aldosteronism pathophysiology manifests across an entire distribution. And it shows that primary aldosteronism is not a categorical or binary disease. You cannot identify on this figure where primary aldosteronism begins. So on the flip side of the coin, it shows the limitations or pitfalls of a categorical approach to designating primary aldosteronism. The categorical definition of a positive oral sodium suppression test is 12 micrograms of aldosterone produced over a 24-hour period. I think that's been around for maybe 50-ish years, and it's relatively arbitrary. If I superimpose this line, you can see that it bisects this continuous distribution. And by doing so, it implies that anybody over this red dashed line has primary aldosteronism. And it simultaneously implies that anyone below this red dashed line does not have primary aldosteronism. But the people below this line also have primary aldosteronism pathophysiology. It's just milder. Okay, so for the sake of argument, let's take this categorical line. What's the estimated prevalence? It would be 9% in normotensive people, 15 to 20% in hypertension, and a quarter of all patients with resistant hypertension. Now again, you can modulate this prevalence. Use a more conservative or more liberal cutoff, and you'll inflate or decrease these numbers. One other thing I'll point out is in individuals with resistant hypertension who are confirmed to have primary aldosteronism, on the day of testing, a large proportion, almost a quarter, had a plasma aldosterone value that was below 10 nanograms per deciliter. And I'll just point this out because many clinicians think that this cutoff for an aldosterone below which excludes the possibility of primary aldosteronism. And this shows the difference between a spot plasma aldosterone value and a 24-hour urinary integrated aldosterone value. Aldosterone production is variable, and a single spot value doesn't tell you the same thing as an integrated value. Even values that you might subjectively think are low may still represent inappropriate pathophysiology. So the prevalence of what I would call categorically overt primary aldosteronism is high. It's mostly unrecognized. One, because we don't look for it. Then when we do look for it, we use relatively arbitrary thresholds that vary even in this room, I'm sure. But beyond this categorical definition of primary aldosteronism, there is a severity spectrum of primary aldosteronism pathophysiology that I showed you ranging from mild to severe. And this continuum exists below our conventional thresholds and is almost completely unrecognized. So the next question to ask is, is this continuum of primary aldosteronism pathophysiology clinically relevant? And if so, at what point does it become clinically relevant? And so to discuss this, I'm gonna switch to epidemiologic data. This is data from large prospective cohort studies, most of which have followed thousands of patients over time. I'll start with normotensive cohort studies. So these are normotensive people on no medications, here seen in this graph, who have a suppressed renin. And if you look on the x-axis, you'll see in the context of this renin suppression, there is a continuum of aldosterone production. And the greater the aldosterone production in the context of renin suppression, the higher the risk of increases in blood pressure over time and the higher the risk of developing hypertension. Similarly, if you're a normotensive person, the greater the magnitude of primary aldosteronism pathophysiology, the greater the incidence of subclinical heart disease, in this case, left ventricular mass index. And this has been seen in several cohorts now, similar findings. Which is to say that if you're normotensive, the magnitude of primary aldosteronism pathophysiology increases the risk of developing hypertension and developing structural heart disease. So what happens when you develop hypertension? Once you develop hypertension, several cohorts from across the world have shown that the magnitude of primary aldosteronism pathophysiology increases the risk of worsening hypertension, hard cardiovascular outcomes, kidney disease, kidney disease progression, end-stage kidney disease, and death. And I think many of you are familiar with a lot of this data, but I'm gonna spend just a minute or two on kidney disease because that has been a very exciting topic in the literature the last couple of years. No doubt, most of you have seen the Fidelio and Figaro studies. These are studies where patients with chronic kidney disease were treated with the novel mineralocorticoid receptor antagonist, venerenone. And when they were treated with this MR antagonist, their progression of kidney disease, develop end-stage kidney disease, and cardiovascular disease outcomes substantially decreased. And the question is, why? And related to this talk, how many of these patients had unrecognized primary aldosteronism that was treated with venerenone? The answer is we don't know because it was not investigated in this study. So we tried to triangulate this in a separate cohort study. This was just a recently completed study that we did in about 4,000 patients with established chronic kidney disease in the CRIC cohort. So among these individuals with chronic kidney disease, as you might imagine, there is a continuum of aldosterone production from relatively low to relatively high. And this parallels a decline in GFR. This is not a surprise. Even though all of these individuals have established kidney disease, the GFR goes down with rising aldosterone. This has been seen before. But as GFR trends down, serum potassium does not increase. It actually tends to trend down. And caloriesis actually tends to increase. Now these are the opposite patterns you would typically see with declining GFR. And they're more in line with pathogenic aldosteronism. So is this pathogenic aldosteronism? Well, if you follow these patients over time, over the next five years, what you find is the greater the aldosterone production in patients with chronic kidney disease, the greater the incidence of CKD progression. That's a 50% decline in GFR. The greater the incidence of end-stage kidney disease requiring dialysis, proteinuria, and death. So there may be many mechanisms by which venerenone benefits these patients. But one interesting postulate is that many of these individuals had unrecognized primary aldosteronism pathophysiology that when treated with venerenone, decreased their risk for disease advancement, both for kidney and heart. So to summarize this epidemiologic data, most of it shows that the magnitude of primary aldosteronism pathophysiology is clinically relevant across the entire continuum. In normotensive individuals, it increases the risk of developing hypertension and structural heart disease. In hypertensive individuals, the magnitude of primary aldosteronism pathophysiology increases the risk of worsening hypertension, incident kidney disease, and heart disease, and death. We discussed that very few of these individuals ever get the categorical definition of primary aldosteronism mainly because we don't look. But of those that do, studies have shown that their risk for kidney disease, heart disease, and death is much higher. And the reason this is so important is we've known for about 50 years, since the early 1970s, that if you treat this biochemical phenotype, renin-independent aldosterone production with a targeted drug like a mineralocorticoid receptor antagonist, even an ENAC inhibitor, you get striking reductions in blood pressure that are then associated with downstream risk reduction. So I will summarize here. I tend to think that primary aldosteronism is not a discrete disease, but rather a syndrome that manifests across a broad spectrum of severity. The prevalence of this primary aldosteronism pathophysiology is very high. It's almost entirely unrecognized. This is clinically relevant because primary aldosteronism pathophysiology contributes to what I would say a large proportion of what we otherwise called idiopathic hypertension, cardiovascular and kidney disease, and we have widely available therapies such as mineralocorticoid receptor antagonists that can mitigate this. So what should we do going forward? I think, as has already been raised, we really need to dedicate ourselves to intensive efforts to increase awareness about primary aldosteronism, testing for primary aldosteronism, not just for countries like the United States and large academic and referral centers, but such that it can be done in all resource settings throughout the world. But simultaneously, we need to think about pragmatic approaches to not just improve the diagnosis, but also implement empiric targeted therapies for presumptive diagnosis. So with that, I will thank, I won't go through individually, but the many mentees, colleagues, collaborators, and funding sources that have contributed to our work over the last many years, and all of you for listening and happy to take questions. Thank you. Let me just fix that. Hello. Rons Redlock from Los Angeles. Very interesting talk. And I wondered if your assessment, based upon the continuous spectrum with less demanding criteria for the diagnosis of primary aldosteronis, is any way changes your concept of the proportion of bilateral to unilateral disease? So none of the data I showed you here directly addresses that question. But I think many studies have already established that the majority of primary aldosteronis is bilateral, and the minority is unilateral. The problem with these assessments is they're heavily biased. They're biased based on who gets tested, as I showed you, is fewer than 1% to 2% who gets referred to imaging and adrenal venous sampling to ultimately make that conclusion, which is a fraction of a fraction of a fraction of that. So of that small fraction of people who get to adrenal venous sampling for subtyping, probably 2 thirds have bilateral disease. If you ask me in the total pie, if you were able to examine it, how many had bilateral disease, I would say probably the vast majority. But we'll probably find those questions better with techniques that Morris Brown described, such as non-invasive imaging, maybe steroid profiling down the road. Yeah. Hello, I'm Emily from London, from Professor Brown's group. Just a question about your continuum data with the urinary aldosterone. Have you broken down your continua according to the patient's demographics, such as their age or ethnicity? We did. So we've previously published that in relation to age, this phenotype, renin-independent aldosterone production, does seem to increase. So the older you are, the higher that phenotype is. With respect to age, sex, race, ethnicity, actually in that publication, we did show that data, except the editors nixed it at the end, which I think was fine. There was no difference by various racial categories by sex. I think what the editors had issue with was this was a heavily biased sample, and not a population-based study, that there was potentially some bias in who participated in those studies. I think that's fair. But what I'll say is, what's not published is, in the spectrum data I showed you, there were no differences by those demographic criteria. OK, thank you. Gomez Sanchez, Mississippi. I remember 50 years ago that we used the value of 12 micrograms as excretion rate of aldosterone. And I didn't know where that came from. So I asked Bob Carey, which is even older than I am, where that come from, and he didn't know either. So I don't know where that come from. But one of the comments I wanted to do is a study that John Lutcher did around 1965, in which he injected, treated aldosterone, and then determined the relative proportion of aldosterone that came as the various metabolites. And one striking thing that has very much importance in your studies is that between 5% to 10% came as the aldosterone 18-oxogluconide. So there's a possibility that many of your patients only had 5%, and therefore you aren't estimating, or you had 10%, and you overestimated the amount of patients that might or might not have abnormal aldosterone secretion. So in order to make it more relevant, perhaps you need to measure other urinary metabolites, like tetrahydroaldosterone, which is 20% to 30%. And perhaps that will give you a much better estimate of how common primary aldosteronism, or inappropriate aldosterone production, might be in your patients. Yes, I love it. I can answer him, right? So, okay, every time I give this talk, I'm always a little bit nervous that somebody who was around 50, 70 years ago will tell me how dare you call these thresholds relatively arbitrary. So I'm glad you said that, because there is no source for 12. Even if it was 10, 10 is just a round number. These are relatively arbitrary, I think you, and Bob Carey's nodding in the back. So if you guys say it's arbitrary, I'm okay saying that. You could tell that I always dance around that, because I don't want to say that these older thresholds are random. But they've been, unfortunately, mentor-to-mentee, generation-to-generation, guideline-to-guideline repeated, so they become the gold standard. And then it's upon our generation to unprove something that was never proven. So I'm glad that you're not discounting that. As to, I have to be brief, we're working on that right now. I completely agree. So in the blood, you're measuring a plasma aldo. Whether you measure it by immunoassay or mass spec, big differences. One moment in time versus 24 hours, big differences. And then the 24 hour is not true 24 hour. It's the acid labile aldosterone, which is probably 1% free, 15% glucuronidated that's freed by acid. So the other 85% is mostly tetrahydraldo. We're working on measuring that now with Rich Aukes, and I think the results are going to look very different. I won't predict which way, but tomorrow we have a presentation that looks into the 18 hybrid steroids as well. And I think you'll get a much better picture if you don't just measure free aldosterone, you try to get the whole metabolite. Thank you very much, Anand. We'll have to kick you off the stage, otherwise. And... APPLAUSE Please, please direct more questions to Anand Arthur. Let's welcome Professor Felix Bollschlein, Professor of Internal Medicine and Endocrinology and Director of the Clinic for Endocrinology, Dermatology and Clinical Nutrition at the University Clinic Zurich in Switzerland. Thank you. Thank you. Professor Bollschlein. Very good. Thanks for having me. I am ending up this particular nice... And that's not... I can do this again, but I would prefer having my own slides. LAUGHTER It was a great talk, though. LAUGHTER What can I do? Not much? Can you please change the slides for Felix Bollschlein? Our last speaker was Felix Bollschlein's talk. This is a talk we already went over. Sorry, Maurice. So we heard much about primary aldo, and this will be more a glance into the future, how you could look at the... Here we go. The disease also from another perspective. And here I was charged with the title of targeted, and I put that in brackets, metabolomics as a tool to discriminate endocrine form of hypertension. I come to this targeted and non-targeted because there are many different ways to look at that. Here we go. Let's try that. Here we go. These are my disclosures. Here's the QR code. And I would like to start with a definition for metabolomics. These are small molecules, large scale of those, that are commonly known as metabolizers in cells, biofluids, tissues or organisms. So we can take this much broader. They are targeted and more untargeted metabolomic approaches. Ones are more focused. They are more sensitive and specific, usually better producible, while the non-targeted approaches are usually those that are for discovery applications. In any way, you have different steps you go through. You have your biological metrics, and these can be starting from different material. You have your data acquisition, data processing. You need a number of statistical analysis. And in the end, you end up with a biomarker discovery, which can be used then for pathway analysis or as a marker for category or non-category disease specification. Now, there are different methods. MRI-based are more likely to be used for untargeted approaches, but not only. And mass spec-based approaches, which more likely are or more often are used for targeted approaches, but again, not only. And the third part that I would also like to introduce, because I'm going to show a number of data, are a combination of those two, in a way where multi-mass spectrometry is used within an imaging approach, so that it's not only the information about the concentration of a given metabolite, but also the position in a tissue, for example, or in a whole organism. And that has been used now already quite heavily, also, here we go, that's a bit, here we go, also in the adrenal gland, and as you can see here, these are normal adrenals. These were studies from Würzburg and Munich. Here's Axel Walsh from the Helmholtz Center in Munich, and you see here very nicely, you can find clusters that represent the zones, zonation, and also the medullary region of the adrenal gland. So that can be used quite nicely, also for micro-dissected tumor samples, as an example, as an intermediate to high-throughput approach, and you can use this then for morphometry, and aphrodisiac e-staining, you can do various approaches for immunochemistry, and then immutable norm analysis. And I'll show you here an example for PBGL. These were 344 tumors, which were collected, and in the next slide, you can see in an unsupervised clustering, here we go, that you can actually make difference between clusters, cluster one and cluster two, as you can see here in this score plot, and of course, this is due to some of the metabolites that you would assume that they are different between different clusters here, DOPA, metoxytyramine sulfate, norepinephrine as examples, but also purine metabolism, or tryptophan metabolism does actually add to this distribution between different clusters. This can be used, not really now necessarily relevant for hypertension, but for categorized PBGL, and for example, metastatic and non-metastatic disease, so you see exanthonuric acid tumor content makes a difference, and this can be then added with the usual clinical risk factors, SDHB mutation, tumor size, and others to categorize tumors. Okay, so how does this relate into, rather, blood samples? This was a study that we did in PBGL patients before and after surgery, comparison baseline with the situation after the tumor had been removed, which makes it easier to compare those samples. These were 56 PBGL patients, 188 metabolites were identified, and then different tests were implied to see what relevant metabolites would come out. And you see here, the main data on the left side, there are all PBGLs included on the right side, more than no adrenergic tumor phenotype, and you see there are differences in metabolites that come out. Above the line, the dotted line, is the increase in change after atrialectomy, or surgical removal, and below the line are metabolites and examples of that change in the opposite direction. So that is a possibility to follow up, and maybe further follow up in PBGL patients. This can also be plotted towards the urine catecholamines, and as many of the clinicians know, hexose or glucose does correlate with catecholamine excess, but so does histidine, as an example, or sphingomyelin, so that this is also a correlation between the endocrine phenotype and effects that might be opposed in target tissues that lead to these changes in the metabolites. This has been done also in another approach, in an untargeted metabolomics, in an NMR-based approach. This was done also in the NSAT-HT consortium, and here you can see that the most significant changes, also in the pre- and post-operative phase, are ketone bodies, again, glucose, organic acids, methanol, dematerialized sulfone, and others. As a bottom line, although the metabolites change from method to method, and from the included patients, from one study to another, there are clear-cut changes that come with the appropriate normalization of catecholamine excess in this instance here. So what about adrenal incidental anoma? So what about cortisol excess? This is a study, again, with the same technique. These are mainly tumors from Irina Bankos, and the same technique, morphometry done, Cp11B1, immunochemistry, and again, the metabolome analysis, and what can you find in the tumors? So first of all, here are the different categories, from non-functional on the left side to overt Cushing syndrome on the right side, according to a text suppression test, and you see this increase, as you would expect, of Cp11B1 expression, the larger, or the higher the endocrine phenotype, the higher the expression pattern. But this can be correlated now with, for example, steroid content. These are mainly adrenal androgens that, as you would expect, with a further suppression of ACTH go down, and you see here, from the left side to the right side, this decrease of, for example, the HEA, cholesterol and sulfate, and other related metabolites. So again, as you would have expected, this is ACTH-driven or non-driven, where you see this decrease. But then there are also other metabolites that you can find, and here is tryptophan metabolism as an example, and you see this increase, really, from non-functional to overt Cushing, with different members of tryptophan metabolism, which goes up. And this is an example for finding metabolites which had otherwise not really been appreciated. And this can be sorted here again in this plot on the above, this CB11B1 expression, very low staining intensity on the left side and to the right, and the high intensity. And you see here the steroids, again, mainly androgens, and you can appreciate this is the reverse picture from the left, higher levels to the right, lower levels, tryptophan metabolism as shown before. But then there are a number of phospholipids that, depending on what phospholipid you look at, you see this converse or parallel increase or decrease with regards to endocrine activity. Okay, so that's, again, tumor metabolism. Looking at blood levels again, we did one earlier study in Cushing's patients, 150 patients from patients where Cushing was excluded up to those with clear-cut overt Cushing syndrome, and we were looking at targeted metabolomics here again. And what did we find is here a number of amino acids, the biogenic amides, which are decreasing with increasing level of cortisol production, and that, of course, is very likely not really related to the tumor itself, but rather the effects of glucocorticoids, for example, on muscle mass, for example, sarcopenia and skin thinning as an example. So it might be a reflection, rather, of the clinical phenotype and the effects on target tissues. Okay, so that can also be used for a discriminant analysis, and on the right side you see some sensitivity, specificity, positive predictive value, and negative predictive value. And, of course, this is a retrospective study, and the numbers are, in some instances, actually quite good, as a retrospective analysis tends to be. Now, switching gears, aldosterone-producing adenomas, we heard already quite a bit about this. Same technique, so here it were 135 patients where the somatic mutation status was present. On the left side, an unsupervised clustering, and you see here above case NG5 as a mutation and COGN1D as another group of patients with good, not perfect, but good separation. And on the right side, again, just an example, that immunochemistry and then the content of steroids can also be measured, and as you would expect, and which has been shown now, of course, in many studies over the years, that 18-oxosteroids are among the best categorizers that can differentiate between tumors that have a case NG5 mutation and those which have not. And this is an example that you really can find those steroids in the tissue, which then go into the blood and can be measured as biomarkers. And that has already been mentioned. The main difference from, and of course, it's a simplification, the difference between aldosterone-producing adenomas and micronodules are the presence or absence of case NG5 mutations, so this mutation and the biomarkers which come with it is essentially, very likely, a good biomarker for unilateral disease, which would be then maybe also for the screening to find those patients where further follow-up and surgical approaches should really be tried. Okay, so, and that is the case in the, this is the prospective study and it's headed by Graeme Eisenhofer, where different steroid metabolites, including the 18-oxosteroids, can distinguish between primary hypertension and primary aldosteronism, but also might in the future help to discriminate unilateral from bilateral diseases, and this is now further looked at in the prospective, in the PROS-ALDO study. Okay, finally, how can this be used? And now this comes really to the question and the title of the talk. Differential diagnosis of hypertension, this is the NSAT-HT Consortium, headed by Maria-Christina Senaro, funded by the EU. There were a number of different metabolites, not really a pragmatic approach, I must say, but really many different techniques are used on the same samples, plasma, urine, steroid metabolomics, metanephrines, microRNA profiles, NMR approaches, and I'm going to show a few data for small metabolites, and I have to say, also, these are still ongoing studies. We don't have all the results yet. There has been a retrospective and a prospective phase, and the data are still in evaluation. These are the retrospective data on the small metabolites. Roughly 300 patients, 280 patients, the same metabolites were measured. A number of data quality analysis were implemented, and then a classical approach and a machine learning approach was implemented to find differences between the groups. And I'm showing you this only data slide here. On the left side, differences between primary hypertension and any form of endocrine hypertension and different metabolites which were found to be significantly different, and on the right side, you see the rock curve which shows how the discrimination between those two groups are. If you look into more detail, into, say, Cushing patients versus essential hypertension, then this differs quite a bit. And this is the last piece of evidence that I'm showing here, so that is the building of a pathway for machine learning approaches where a small metabolite on the left side go in as the data set. The diseases are comparised, then all versus all, or only essential hypertension versus endocrine hypertension. Features are selected. A number of co-variables or co-founders need to be taken into account, and then age and sex is among the most important one. This is then tested and trained to find a classification. And that is to show that this, depending what... Last slide, okay, here we go. To show that you get many data, depending what groups you look at and what data set you look at and what different machine learning approaches you're using that, for example, if you look here in the middle, the dark blue one, essential endocrine hypertension versus primary hypertension, you can, for example, find tools where you have a pretty good sensitivity, which would be a tool to be used, say, in a GP kind of setting, which then further referral to endocrinologist or cardiologist would be a possibility. Okay, finally, so metabolomics do provide a rich source of phenotypic annotations, either from the tumor itself also or other biofluids. They can be used potentially as a screening tool, but also, I think, quite interestingly, to explore disease mechanisms. There are a number of variables that have to be taken into account, and it requires large enough training sets for machine learning-based techniques, and prospective validation is within reach, and I hope, for example, next year, they should be available. So I thank you very much, and I thank many of the contributors, NSUT-HT in Zurich and many other places, and I thank you very much for your attention. Thank you. Thank you for a wonderful talk, Mr. Borshaw. We're now open for questions. Professor Rainey? Rainey from Michigan. Felix, that was a huge amount of work, and it spanned many different techniques and methodologies. I wanted to ask sort of about the last program, and you sort of grouped endocrine hypertension as this homogenous group. Obviously, there's multiple causes. Are you able to break the endocrine hypertension out into the actual causes, Cushing, PA, VO? So it's really a question how you look at the data, and of course, a great tool would be that you have one blood sample, and it says, well, that's PBGL, that's Cushing, that's primary aldo. Obviously, you can do the cystic suppression tests and L-urenine ratio and metanephrine measurements and so on, but it's possible. If you look rather between primary hypertension and endocrine forms of hypertension, and again, yes, there are a number of categories which are completely different, but they are metabolized, which actually go in the same direction, and it's a matter of how you look at this statistically or from these machine learning tools which are agnostic for that. So I think it's possible to have larger categories such as endocrine hypertension, and probably the same is true for other forms of secondary hypertension, but then, obviously, yes, the different subcategories can be further looked at, but then you need larger numbers. In the prospective phase of NZHT, there are now roughly 1,500 patients who have been very good characterized, and I think these numbers will be large enough to do that then. Hi, Felix, that was great. My question actually builds on Bill's and what you were just saying, which is it really depends on how you analyze the data. So, of course, this is my bias, of course, that a lot of these analyses categorize primary hypertension and a form of secondary hypertension, and it seems like you have an opportunity to ask a deeper question, which is what is primary hypertension? Presumptively, it's the exclusion of other secondary forms, and even what is secondary hypertension? So have you considered doing, for example, a waterfall plot of one metabolite or a combo of metabolites? My guess is what you would see is primary hypertension and secondary hypertension get blurred, and you may be able to distinguish severe extremes, but also find the blur in the middle, and that could, you know, because otherwise, I think you suggested this, external validation will keep giving you the same answer since everybody's diagnosing these things in a dichotomous way. Absolutely, and I think you, I mean, you showed the non-categorizing aldosterone output in the kidney, and I would assume, and actually I think that we even started looking into this, that it's probably true for pretty much all the metabolites. It's hard to think that there's a metabolite which is zero and goes to one, and there's nothing in between. There will always be gray zones, or you can actually use the gray zone for sub-categorization of the essential hypertension. So actually, among the future plans is to use these kind of fingerprints to look into an unselected group of hypertensive patients to say or to make the hypothesis that there are subgroups of essential hypertension which behave rather like, you know, hypercortisolism, or rather like adrenergic kind of phenotype, or aldosterone-driven phenotype, and maybe even use this information for a targeted medication, this MRI, this beta-blockage, the ACE inhibitors, and so on. So that is, we hope we get funded. That's the NSAT-HT plus kind of study. But that would be among the ways to look at, to sub-categorize patients with hypertension in some endocrine kind of flavors, and use this information for treatment. Great, thank you. Is there someone at the back? Is there someone at the back wanting to ask a question? If not, there's just one more question from online. Could you use these markers to predict or evaluate response to particular therapies? Well, that goes in a similar direction, not yet, and the study would look like that the multi-omics would be done, and then a treatment round, because you would require, of course, the information on treatment response, so you have a round of beta-blocker, ACE inhibitor, MRI, and so on, and then you would define treatment response, which again is a tricky issue, I would say, but then you would try to categorize treatment responders towards this metabolite pattern, and make in a further prospective study then the prediction that this pattern is predicting treatment response. So, don't know yet. Thank you. Thank you very much, and please put our hands together for a great lineup of speakers in this session, and we'll close the session. Thank you.

primary aldosterism

aldosterone

mineralocorticoid receptor

kidney

high blood pressure

cardiovascular diseases

diagnostic approaches

endocrine hypertension

resistant hypertension

genetic basis

genotype-guided therapy

metabolic changes

screening tool