Talks and tweets from the ADA

Every year at the American Diabetes Association (ADA) annual meeting, thousands of presentations are made. This year's 78th Scientific Session took place from 22 to 26 June in Orlando, Florida. I have attended (nearly) every year from 1978, and wrote several hundred articles summarizing various sessions over nearly two decades in the “Perspectives on the News” column in Diabetes Care. How, I wondered, could I give the readers of the Journal of Diabetes a sense of the fascinating breadth of information and understanding gleaned over the 5-day period? Here's an approach: a summary of my @zbloomgarden “tweets” of posters and of my notes from the meeting, recognizing that any such summary must be woefully incomplete. (Note, abstracts referred to below [numbers in parentheses] can be viewed at https://plan.core-apps.com/tristar_ada18/abstracts [accessed 3 August 2018]; many actual posters are available at https://ada.scientificposters.com/epsWelcome.cfm [accessed 3 August 2018]). Barbara Corkey (Boston University, Boston, MA, USA) made the important point that although diabetes is the end result of multiple pathogenic abnormalities, it is incorrect to consider obesity, hypertension, insulin resistance, hyperinsulinemia, decreased insulin clearance, dyslipidemia, inflammation, non-alcoholic fatty liver disease, or any one of the myriad of interrelated factors as being “primary”; rather, one should think of all these factors as having the potential to “cause each other”. And, bringing together diabetes pathogenesis and cardiovascular disease (CVD), Lim et al. (Abstract 455) found that among >1000 people with diabetes, the levels of branched-chain amino acids (which have been shown to track with insulin resistance) predict subsequent heart failure. In an interesting study (Abstract 296-LB), leptin levels were measured in 1064 obese people, and the leptin analog metreleptin was administered to those in the lowest leptin tertile. Among this lower third, those with the lowest baseline leptin had substantial weight loss, suggesting that in a (small) subset of people with obesity this may be a reasonable therapeutic approach. In a perhaps more practical study (Abstract 298-LB), a meal replacement approach was more effective than a food-based diet over 1 year, with 10.3% vs 5.5% weight loss and with diabetes resolution in approximately one-third versus one-fifth of subjects. There may be a genetic basis to glucose–HbA1c mismatch, with some people being “high glycators” and others “low glycators”, a not uncommon issue in diabetes (Abstract 183-LB). Bernard Zinman (Mount Sinai Hospital, Toronto, Canada) addressed this in a symposium on diabetic kidney disease, first reviewing evidence from Diabetes Control and Complications Trial (DCCT)/Epidemiology of Diabetes Interventions and Complications (EDIC)1, 2 and from Action in Diabetes and Vascular Disease: Preterax and Diamicron MR Controlled Evaluation (ADVANCE)3, 4 that better glycemic control particularly tracks with improved renal outcome over the long-term in post-trial follow-up. HbA1c (affected by anemia and renal insufficiency), fructosamine (affected by hemoglobin, uric acid, and bilirubin), and glycated albumin (affected by hypoalbuminemia and proteinuria) are ultimately indirect approaches, suggesting that continuous glucose monitoring may be the optimal approach in determining overall glycemia as well as hypoglycemia frequency. Edgar Peters (VU University Medical Center, Amsterdam, Netherlands) spoke at a symposium on the diabetic foot, asking whether 6 weeks is a sufficient duration of antibiotic treatment for osteomyelitis, and noting that an elevation in the sedimentation rate may persist for much longer periods, and may serve as a more useful biomarker than C-reactive protein (CRP) and procalcitonin in guiding the duration of antibiotic treatment, given the potential for adverse outcome associated with recurrence after antibiotic discontinuation. Two fascinating debates at the ADA addressed aspects of diabetic nephropathy. Peter Rossing (Steno Diabetes Center, Gentofte, Denmark) argued for and Robert Nelson (National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Phoenix, AZ, USA) argued against the question of whether albuminuria in the absence of hypertension should be treated with angiotensin-converting enzyme inhibitors (ACEi) or angiotensin receptor blockers (ARB). Rossing pointed out that it is clear that such treatment is of benefit with blood pressure (BP) >140/90 mmHg, and noted that the time from microalbuminuria to end-stage renal disease (ESRD) may exceed 15–20 years, beyond the duration of existing clinical trials, leading to the question of what constitutes a normal BP, what situations should be considered high risk, for which 130/80 mmHg is the current ADA goal, and whether a normal BP may constitute masked hypertension under certain circumstances, further noting that BP lowering does not track well with the albuminuria-lowering effect of ACEi/ARB. Nelson suggested that relevant outcomes are renal failure, CVD, and cardiovascular (CV) mortality, but not microalbuminuria, pointing out that microalbuminuria may regress in association with lower BP, but also with lower levels of glycemia and lipids, and noting studies in which ACEi and ARB did not lead to improvement in albuminuria, and even in which ARB were associated with worse renal outcome, arguing that until biomarkers are found to show which patients with microalbuminuria benefit from treatment he would not use such agents. Dick de Zeeuw (University of Groningen, Groningen, The Netherlands) argued that albuminuria should be considered a therapeutic target, noting its association with endothelial dysfunction and its independent association with both CV and renal outcomes, and pointing to the relationship between albuminuria and endothelial glycocalyx, with evidence that sulodexide restores glycocalyx and reduces albuminuria. However, Merlin Thomas (The University of Melbourne, Melbourne, Vic., Australia) espoused the viewpoint that “the target is never albuminuria”, a surrogate that one would not expect to itself be a determinant of the benefit of a given treatment approach, making points that albuminuria is more of a symptom than a cause of renal disease, that the evidence shows benefits of ACEi/ARB rather than of specific albuminuria targets, and that there is at least the concern that maximally lowering albuminuria (e.g. with dual ACEi and ARB, or even with these plus a mineralocorticoid antagonist) may have a safety risk of increasing the likelihood of worsening renal function, particularly in people whose baseline renal function is severely compromised. The excitement of the recognition that sodium–glucose cotransporter 2 (SGLT2) inhibitors, glucagon-like peptide 1 receptor agonists (GLP1-RA), and thiazolidinediones are associated with CV benefit in diabetic people with existing CVD has led to new concepts of treatment. As Sanjay Kaul (Cedars-Sinai Medical Center, Los Angeles, CA, USA) commented, it would be hard to imagine “going back” to introducing new diabetes treatment approaches without CV outcome trials (CVOTs). Such trials in people with existing CVD require approximately 15 000 person-years of exposure. There is now recognition of heart failure, as well as atrial fibrillation and sudden death, as important additional CV outcomes, and there is intriguing evidence of potential for nephroprotection with SGLT2 inhibitors, as well as perhaps with incretin-based treatments. Kaul also noted, “we really have yet to study the low-risk population”, but there is a major issue: because of their much lower event rates, CVOTs for all people with type 2 diabetes mellitus (T2DM), rather than just those with existing CVD, may need more than a fourfold increase in numbers of enrollees, and following tens of thousands of people with diabetes for a decade would be a massive undertaking. An interesting suggestion made by Steven Marso (University of Texas Southwestern Medical Center, Dallas, TX, USA) was to consider enrollment of patients with a positive coronary calcium score, a group at intermediary risk. Other markers for such studies may include microalbuminuria and elevation in CRP. A fascinating subgroup analysis of the Exenatide Study of Cardiovascular Event Lowering (EXSCEL) Trial of once weekly exenatide used propensity matching to assess CV outcomes among those of the 7396 placebo-treated participants who received an SGLT2 inhibitor compared with those not so treated, finding a 21% lower rate of major adverse cardiovascular events (MACE), 50% lower all-cause mortality, and significant improvement rather than reduction in estimated glomerular filtration rate (eGFR; Abstract 130-LB). An implication: if some 10% of those receiving placebo but fewer of those receiving exenatide received these agents (additional treatment presumably being given for those needing better glycemic control), then part of any CV benefit of exenatide q.w. may have been masked. At a symposium on digital approaches to implementing knowledge from the National Diabetes Education Program, Ann Albright (Centers for Disease Control and Prevention, Atlanta, GA, USA), Linda Siminerio (University of Pittsburgh, Pittsburgh, PA, USA), John Piette (University of Michigan, Ann Arbor, MI, USA), and Athena Philis-Tsmikas (The Whittier Institute for Diabetes, La Jolla, CA, USA) reflected on various digital approaches, representing a continuum from mobile technology, texting, and emailing to a full “virtual lifestyle change program” based on “wearables”, apps, and coaches. Such approaches appear to have been effective in open-label uncontrolled studies of patients with diabetes in poor control, but there are issues with adherence and with real-world implementation, particularly in integration with electronic medical records and in making the intervention meaningful to the patient. Specific interventions seem possible, such as assuring that a population of people with diabetes has retinal screening or foot examination, but texting medication and blood glucose self-monitoring reminders had limited long-term efficacy. Judith Fradkin (Division of Diabetes, Endocrinology, and Metabolic Diseases, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD, USA) pointed out the interest of the National Institute of Diabetes and Digestive and Kidney Diseases in digital translational research, noting that funding is available for appropriate products (see https://grants.nih.gov/grants/guide/pa-files/PAR-17-178.html, accessed 3 August 2018). An interesting study showed that adding saxagliptin to dapagliflozin plus metformin substantially reduced genital tract fungal infection from approximately 5% to 3% of treated people, perhaps an approach for people with this side effect, although one that may simply be due to reduction in glycosuria because of lower glucose levels with the addition of the DPP-4 inhibitor (Abstract 1171). As a reminder that even minor degrees of renal insufficiency attenuate the glucose-lowering effect of SGLT2 inhibitors, a comparison of sitagliptin versus dapagliflozin showed HbA1c reductions of 0.51% versus 0.36% at baseline eGFR of 78 mL/min per 1.73 m2 and HbA1c 7.8% (Abstract 1142). Sitagliptin also showed benefit when continued after initiation of insulin treatment, leading to a 1.88% reduction in HbA1c, whereas administration of insulin with placebo led to a 1.42% reduction (Abstract 112-LB). A network meta-analysis of drugs given in combination with metformin showed that DPP-4 inhibitors, thiazolidinediones, and GLP-1RA were associated with 27%, 22%, and 25% lower CVD mortality than sulfonylureas, whereas neither meglitinides nor α-glucosidase inhibitors were associated with better outcome (Abstract 1239). A propensity score matching comparison of 7561 people receiving DPP-4 inhibitors with the same number receiving a sulfonylurea showed a 34% reduction in the likelihood of dementia (Abstract 195-LB). A propensity score-matched comparison of dapagliflozin with a DPP-4 inhibitor in the CVD-REAL2 study showed the former to be associated with 26%, 13%, 14%, and 18% lower likelihood of death, heart failure, myocardial infarction, and stroke, respectively (Abstract 124-LB). A 1-year study of 1550 people with type 1 diabetes mellitus (T1DM) receiving placebo or the SGLT1/2 inhibitor sotagliflozin at daily doses of 200 or 400 mg showed less severe hypoglycemia (7.4% vs 5.7% and 4.4%) and greater reduction in HbA1c (by 0.2% and 0.3%), but diabetic ketoacidosis (DKA) in 0.2% vs 2.9% and 3.8% (Abstract 5-LB). The use of SGLT2 inhibitors in T1DM was the topic of another symposium at the ADA, again showing greater DKA occurrence with dapagliflozin, with empagliflozin (apparently not with a very low 2.5-mg dose), and with the dual SGLT1/2 inhibitor sotagliflozin. A recent meta-analysis of 11 trials of 3523 T1DM people treated with an SGLT2 inhibitor versus placebo failed to show a significant reduction in severe or total hypoglycemia, but did show a 0.4% reduction in HbA1c and a 2.34-fold increase in DKA, occurring in 1.54% of those receiving the SGLT2 inhibitor compared with 0.46% of those receiving placebo.5 One must wonder whether this in itself offers a sufficiently great benefit: risk ratio. A more compelling possibility was raised by David Cherney (University of Toronto, Toronto, Canada): might a cardiovascular/renal prevention trial with an SGLT inhibitor in T1DM, improve the “tremendous morbidity/mortality [and] unmet need” in these patients? A 1-year study of 26 people with T1DM receiving liraglutide versus 20 receiving placebo showed a 0.5%–0.6% HbA1c reduction from a baseline of 7.8%, in association with weight loss and BP lowering, but without reduction in hypoglycemia (Abstract 3-LB). A comparison of insulin patch pumps with insulin pen treatment using mulitple daily injections showed no difference in glycemic response on continuous glucose monitoring (Abstract 73-LB). A comparison of a novel rapidly absorbed insulin aspart preparation with the currently used preparation in 1445 diabetic people showed a 16% reduction in nocturnal hypoglycemia, but no difference in daytime events (Abstract 96-LB). A new glucagon preparation administered by nasal spray was may soon be available (Abstract 138-LB), offering greater potential usefulness than the existing glucagon emergency kits. The ADA meeting ended with a symposium in which the writing group presented the 2018 ADA-EASD consensus report on therapeutic approaches to hyperglycemia in T2D. Judith Fradkin noted the benefits of glycemic control in averting symptomatic hyperglycemia, with evidence of a reduction in microvascular complications by 50%–75% with HbA1c of 7% vs 9% in the DCCT and by 25% with HbA1c of 7% vs 7.9% in the UK Prospective Diabetes Study (UKPDS), but with greater benefit with reduction from higher levels of HbA1c. She pointed out that there is, however, uncertainty regarding macrovascular benefits of glycemic control in T2DM, and that these benefits appear to emerge slowly while adverse effects can be seen earlier, necessitating that risks and benefits be balanced, considering patient preferences. She noted that although HbA1c is the major tool used in assessment of glycemic control, with anemia or chronic kidney disease there may be discrepancies between HbA1c and glycemia, and that self-monitoring of blood glucose (SMBG) is useful for self-management with insulin treatment, although of more limited benefit outside insulin treatment with additional cost, while she considered the role of newer technologies not to have been established. Deborah Wexler (Massachusetts General Hospital, Boston, MA, USA) introduced the truly novel aspect of the consensus report, the recommendation that the presence of CVD is the major determinant of treatment recommendation: GLP-1RA are favored if atherosclerotic CVD predominates, in the order of liraglutide > semaglutide > exenatide q.w., or considering the SGLT2 inhibitor empagliflozin (interestingly, pioglitazone was not mentioned here despite the evidence of benefit in CVOTs); and SGLT2 inhibitors are recommended if heart failure predominates, and if the eGFR is adequate, with the recommendation to use GLP-1RA at lower eGFR levels, while avoiding thiazolidinediones. The potential renal benefits of SGLT2 inhibitors and liraglutide (based on its effect on albuminuria) were mentioned. Walter Kernan (Yale University, New Haven, CT, USA) further reviewed therapeutic options, mentioning among other points that long-acting insulin preparations may be associated with less hypoglycemia than neutral protamine Hagedorn (NPH) insulin, but that the former may not be better for all patients. Geltrude Mingrone (Catholic University, Rome, Italy) noted several disadvantages of SGLT2 inhibitors, in particular the increase in low-density lipoprotein cholesterol and the increase in amputation rates with canagliflozin and ertugliflozin, and opined that GLP-1RA “probably do not increase risk for pancreatitis…nor pancreatic cancer at least in the short term”. David D'Allessio (Duke University, Durham, NC, USA) gave the committee's recommendation “that metformin remain the initial drug…but there is the new data and new alternatives…[although] the safety, efficacy and costs continue to weigh”. He suggested considering initial combination treatment if a > 1% reduction in HbA1c is desired, reminded the audience that CV risk “be considered upfront” but “for the other 80%…need to minimize hypoglycemia…[and consider] issues of body weight”, with issues of cost leading to the use of metformin, then a sulfonylurea and/or a thiazolidinedione, in either order. Chantal Mathieu (University Hospital Gasthuisberg, Leuven, Belgium) opined that “the evidence now is strong enough” for combined use of GLP-1RA and insulin, whether as separate agents or in fixed-ratio premixed formulations, with consideration to adding an SGLT2 inhibitor, recognizing the risk of DKA with overly aggressive insulin down-titration and in the setting of surgery and other stress situations. John Buse (University of North Carolina, Chapel Hill, NC, USA) ended the session with some “knowledge gaps”, in developing approaches to personalization of treatment, in considering whether metformin's role as foundational treatment is just “a quirk of history”, in the use of SGLT2 inhibitors and GLP-1RA in primary prevention, and in whether they may have additive CVD benefit, in determining the appropriate use of SMBG and of continuous glucose monitoring, and a myriad of other fascinating questions. 每年在美国糖尿病协会(American Diabetes Association,ADA)的年会上都有数以千计的报告。今年第78届ADA科学大会于6月22日至26日在佛罗里达州奥兰多举行。自从1978年以来我(几乎)每年都会参会, 并且近二十年在Diabetes Care的《新闻透视》专栏中撰写了数百篇总结各种会议的文章。我在这5天的时间里收集到的海量信息和思考, 怎样才能够使得Journal of Diabetes的读者对此有所了解?在这里有一种方法:我的@zbloomgarden账号发送的“推文”的海报总结以及我的会议纪要, 但是我们一定要理解任何此类总结都不可能是完整的(注释:在https://plan.core-apps.com/tristar_ ada18/abstracts网站可以查看以下提及的摘要信息[括号中的数字] [accessed 3 August 2018];在https://ada.scientificposters.com/epsWelcome.cfm网站可以查看到许多原先展示的海报[accessed 3 August 2018])。 Barbara Corkey(Boston University, Boston, MA, USA)曾经提出了一个重要的观点, 即糖尿病虽然是多种异常致病因素的最终结果, 但是如果将肥胖、高血压、胰岛素抵抗、高胰岛素血症、胰岛素清除率下降、血脂异常、炎症、非酒精性脂肪性肝病或者众多相关因素中的任何一种当成是“主要致病因素”却是一种错误的观点;相反, 我们应该考虑到所有这些因素都有可能“彼此造成影响”。而且,Lim等(摘要455)在> 1000名的糖尿病患者中调查后发现, 糖尿病的发病机制与心血管疾病(CVD)相关, 支链氨基酸水平(既往已经证实它与胰岛素抵抗相关)可以预测随后发生的心力衰竭。 在一项有趣的研究(摘要296-LB)中测量了1064名肥胖患者的瘦素水平, 并且对瘦素水平处于最低三分位数组的患者给予瘦素类似物美曲普汀治疗。在这个最低三分位数组中, 基线瘦素水平最低的患者体重下降非常明显, 这意味着在这个(小型)肥胖亚组人群中使用美曲普汀治疗可能是一种合理的治疗方法。在一项也许更为实际的研究(摘要298-LB)中, 经过1年以上的观察后发现膳食替代疗法与基于饮食控制的疗法相比更为有效, 两组受试者的体重分别下降了10.3%与5.5%,糖尿病缓解率大约分别为三分之一与五分之一。 患者的葡萄糖-HbA1c不相匹配可能有其遗传基础, 有些人容易“高糖基化”, 而有些人容易“低糖基化”, 这在糖尿病患者中并非罕见(摘要183-LB)。Bernard Zinman(Mount Sinai Hospital, Toronto, Canada)在糖尿病肾病专题研讨会上对此做了一个发言, 他首先回顾了来自于糖尿病控制与并发症试验(Diabetes Control and Complications Trial,DCCT)/糖尿病干预与并发症流行病学调查研究(Epidemiology of Diabetes Interventions Complications,EDIC)1,2以及来自于控制糖尿病与血管疾病行动研究:百普乐与达美康缓释片对照评估试验(Action in Diabetes and Vascular Disease: Preterax and Diamicron MR Controlled Evaluation,ADVANCE)3,4的证据, 在这些试验之后的长期随访中发现, 更好的血糖控制尤其有利于改善肾脏结果。而控制HbA1c(受到了贫血与肾功能不全的影响)、果糖胺(受到了血红蛋白、尿酸以及胆红素的影响)以及糖化白蛋白(受到了低蛋白血症与蛋白尿的影响)最终都是间接的方法, 这意味着连续血糖监测可能是用来测定整体血糖控制情况以及低血糖频率的最佳方法。 并发症 Edgar Peters(VU University Medical Center, Amsterdam, Netherlands))在糖尿病足专题研讨会上发表了一个讲话, 关于在使用抗生素治疗骨髓炎时, 考虑到停用抗生素可能导致与复发相关的潜在不良后果, 提问6周的疗程是否充足, 并且指出了红细胞沉降率的升高可能会持续更长时间, 与C-反应蛋白以及降钙素原相比可能是更好的指导抗生素治疗时间的生物标志物。 Diabetes Center, Gentofte, Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Phoenix, AZ, mmHg, de of Groningen, Groningen, The University of Melbourne, Melbourne, Vic., Medical Center, Los Angeles, CA, outcome of Texas Southwestern Medical Center, Dallas, TX, Study of Cardiovascular Event Diabetes Education Ann for Disease Control and Prevention, Atlanta, GA, of Pittsburgh, Pittsburgh, PA, of Michigan, Ann Arbor, MI, Whittier Institute for Diabetes, La Jolla, CA, of Diabetes, Endocrinology, and Metabolic Diseases, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD, accessed 3 August mg of Toronto, Toronto, Prospective Diabetes General Hospital, Boston, MA, University, New Haven, CT, University, Rome, University, Durham, NC, Hospital Gasthuisberg, Leuven, of North Carolina, Chapel Hill, NC,