Tweeting, tweeting.

It is a useful exercise in brevity to limit one's thoughts to 280 characters and an image. I've been “tweeting” for more than a decade, and in 2018 used tweets for the basis of an American Diabetes Association meeting report. In the era of COVID‐19, we are moving to virtual patient visits and virtual meetings, and online forums have the potential to enrich our understanding of diabetes. Here's a look over my 2020 Twitter notes.

In the present issue of the Journal of Diabetes, Jing et al show that low bone mineral density in 482 men with type 2 diabetes tracked with low estradiol and high follicle‐stimulating hormone, without effect of testosterone or sex hormone‐binding globulin. This implies an important role of metabolism and sex hormone changes in bone homeostasis among people with diabetes. Clinical questions about nonalcoholic fatty liver disease (NAFLD) abound. One is whether incidentally discovered NAFLD on imaging has the same consequence as NAFLD associated with elevations in liver function tests. In a report of a ≥7‐year follow‐up of imaging and alanine aminotransferase (ALT) measurements from 130 US Veteran's Administration hospitals from 2004 to 2008, 15 419 patients had NAFLD and elevated ALT, and 9267 patients had neither NAFLD nor elevated ALT; the former group had approximately 4‐fold greater likelihood of developing cirrhosis. An additional 3522 patients had NAFLD and normal ALT, and showed no significant increase in cirrhosis. A series of publications in Diabetologia in 2009 raised concern that insulin glargine was associated with increased likelihood of cancer risk, although the data never seemed sturdy enough to avoid use of this important basal insulin preparation.4, 5, 6 Now a dataset study of more than 300 000 insulin‐treated women initiating treatment with glargine, detemir, or neutral protamine Hagedorn (NPH) insulin appears to have permanently put to rest the notion that insulin glargine might have an association with breast cancer.

Precision management of type 2 diabetes now may include the use of genetic information. Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial participants with the haptoglobin (Hp) 2‐2 genotype had reduction in atherosclerotic cardiovascular disease endpoints with intensive glycemic treatment, while Hp 1‐1 and Hp 1‐2 not only were associated with lack of CVD benefit but also with increased mortality when subjected to intensive treatment. Another genotype/phenotype analysis of ACCORD showed that the T/T PPARα genotype was associated with nearly a halving of major adverse cardiac events (MACEs) with fenofibrate treatment. This is a CVD outcome benefit similar to that seen in participants both with high‐density lipoprotein cholesterol <35 and triglyceride >203 mg/dL. Another potentially important genetic heterogeneity affecting sodium glucose cotransporter 2 inhibitor (SGLT2i) response, although not readily ascertained in clinical management: Variable expression of alpha cell SGLT2 RNA and protein appears to track with variation in dapagliflozin‐induced in vitro glucagon secretion from 31 islet donors.

Recent population studies showed an additive effect of chronic kidney disease (CKD) and diabetes on postmyocardial infarction mortality, and an additive effect of diabetes and hypertension control both on albuminuria and on left ventricular hypertrophy, to an extent similar to that reported in ACCORD and other trials. A trial of hypertension treatment at mean age 81 showed that reducing systolic pressure to 128 vs 144 improved brain change on MRI, as well as being associated with reduction in CVD outcome, supporting a systolic blood pressure goal of 130 regardless of age. What antihypertensive treatment is used is however important. Among more than 700 000 hypertensive persons treated with chlorthalidone or hydrochlorothiazide, CVD risk was similar, but there was greater likelihood of harm with the former agent, which was associated with greater likelihood of type 2 diabetes, acute renal failure, CKD, hypokalemia, hypomagnesemia, and hyponatremia, and also with hyperkalemia, anaphylactoid reactions, and gout. A study of hypomagnesemia showed it to be associated with increased CVD mortality among more than 3000 persons with CVD followed for >7 years. A recent subgroup analysis of the dapagliflozin cardiovascular outcome trial (DECLARE‐TIMI58) showed that with a diabetes duration of >20 years, the agent not only reduced heart failure but also MACEs and all MACE components. In a report on 3909 ACEI/ARB (angiotensin‐converting enzyme inhibitor/angiotensin receptor blocker)‐treated persons whose estimated glomerular filtration rate decreased to <30 mL/min, 1235 discontinued the treatment within 6 months while 2674 did not. Over nearly 3 years, propensity score‐matched mortality and MACE risk increased 39% and 37% in those stopping ACEI/ARB treatment. At the beginning of the natural history of diabetes, a meta‐analysis showed reduction in risk of diabetes development in randomized controlled trials with ACEIs, and a Mendelian randomization analysis showed association of genetically lower serum ACE levels with lower diabetes risk, perhaps in part mediated by lower body weight with lower ACE level. One wonders whether measurement of circulating ACE levels would allow greater precision in determining persons appropriate for ACEI/ARB treatment.

Two other noteworthy reports: Analysis of >17 000 diabetic ketoacidosis hospitalizations in Australia and New Zealand showed serum osmolarity to have a greater effect than the degree of acidosis in predicting adverse outcome. Neuropathic pain is a major issue in clinical management, but an issue is the potential of pharmacologic management to have significant side effects, including sedation and weight gain. We have reviewed studies showing benefit of acupuncture, and the recent publication of a trial comparing real with sham acupuncture found an impressive halving of neuropathic pain with the former.

将一个人的想法限制在280个字符和一张图片之内，这是一个简洁而有用的练习。十多年来，我一直在发“推特”，2018年，我以推特为基础在美国糖尿病协会做了一份会议报告。在COVID‐19的时代，我们正在转向虚拟患者探视和虚拟会议，这些在线讨论的方式有可能丰富我们对糖尿病的理解。下面是我在2020年的一些推特笔记。

在最新一期的Journal of Diabetes上Jing等人的研究表明，482名2型糖尿病男性的低骨密度与低雌二醇和高卵泡刺激素有关，而不受睾酮或性激素结合球蛋白的影响。这意味着代谢和性激素变化在糖尿病患者的骨稳态中起着重要作用。关于非酒精性脂肪性肝病(NAFLD)的临床问题比比皆是。在影像学上偶然发现的NAFLD是否与肝功能检查中有异常升高的NAFLD具有相同的预后就是其中一个例子。 在2004年至2008年间对130家美国退伍军人管理局医院进行的成像和丙氨酸氨基转移酶(ALT)数据长达7年的随访报告中，15419名患者患有NAFLD且ALT升高；9267名患者既没有NAFLD，也没有ALT升高；前者发展为肝硬化的可能性大约高4倍。另有3522名患者患有NAFLD但ALT正常，没有显示出肝硬化的显著增加。2009年在Diabetologia杂志上的一系列文章引起了人们的关注，即甘精胰岛素与癌症风险的增加有关，然而数据并未有说服力到避免使用这种重要的基础胰岛素制剂。而现在，一项使用甘精、地特或中性鱼精蛋白Hagedorn(NPH)胰岛素治疗300000多名女性的研究似乎永久地打消了甘精胰岛素可能与乳腺癌有关的想法

现在对2型糖尿病的精确管理可能需要使用到遗传信息。一项旨在控制糖尿病心血管风险的行动(ACCORD)试验发现，具有结合珠蛋白(HP)2‐2基因型的参与者在强化血糖治疗时，动脉粥样硬化性心血管疾病的发生减少，而Hp1‐1以及Hp1‐2不仅对心血管缺乏益处有关，而且与在接受强化治疗时死亡率增加有关。ACCORD的另一项基因型/表型分析显示，T/T PPARα基因型可以降低几乎一半的非诺贝特治疗导致的主要不良心脏事件(MACEs)。这与高密度脂蛋白胆固醇<35 mg/dL和甘油三酯>203 mg/dL的参与者心血管结局益处相似。此外，尽管在临床治疗中并不容易确定，一个影响钠葡萄糖共转运体2抑制剂(SGLT2i)反应的潜在重要遗传异质性，α细胞SGLT2 RNA和蛋白的可变表达可能与达格列净诱导的31名胰岛捐赠者体外胰高血糖素分泌的变化有关。

最近的人群研究表明，慢性肾脏疾病(CKD)和糖尿病对心梗后死亡率有加成作用，控制糖尿病和高血压对治疗蛋白尿和左心室肥厚也有加成作用，在一定程度上与ACCORD和其他试验报告的程度相似。一项在平均年龄81岁时进行的高血压治疗试验表明，将收缩压降低到128mmHg时相对144mmHg可以改善MRI上的脑部变化，并与降低心血管疾病的结局有关，因此，无论年龄如何，收缩压目标都应该控制在130mmHg。然而，使用什么样的降压治疗是很重要的。在70万多名接受氯噻酮或氢氯噻嗪治疗的高血压患者中，他们心血管疾病的风险相似，但前者的伤害可能性更大，它与2型糖尿病、急性肾功能衰竭、慢性肾脏病、低钾血症、低镁血症和低钠血症，以及高钾血症、类过敏反应和痛风更有关。一项关于低镁血症的研究表明，在3000多名随访超过7年的心血管病患者中，低镁血症与心血管病死亡率的增加有关。最近对达格列净心血管结局试验(DECLARE‐TIMI58)的亚组分析显示，糖尿病病程>20年时，该药不仅减少了心力衰竭，还减少了MACEs和所有MACE事件。在一份关于3909名接受ACEI/ARB(血管紧张素转换酶抑制剂/血管紧张素受体阻滞剂)治疗的肾小球滤过率降至<30mL/min的患者的报告中，1235人在6个月内停止治疗，另外2674人继续治疗。在近3年的时间里，那些停止ACEI/ARB治疗患者的倾向得分匹配死亡率和MACE风险分别增加了39%和37%。在糖尿病自然病史开始时，meta分析显示，在使用ACEI的随机对照试验中，糖尿病发生的风险降低，孟德尔随机化分析显示，遗传上较低的血清ACE水平与较低的糖尿病风险有关，可能部分是由于较低的体重和较低的ACE水平介导的。那么不禁让人思考检测循环中ACE的水平是否可以更精确地确定哪些患者更适合ACEI/ARB治疗。

另外两个值得注意的报告是:对澳大利亚和新西兰超过17000例糖尿病酮症酸中毒住院患者的分析显示，血清渗透压在预测不良结果方面比酸中毒的程度有更大的作用。神经性疼痛是临床治疗中的一个主要问题，但是药物治疗可能会产生显著的副作用，包括镇静和体重增加。我们回顾了关于针灸益处的研究，最近发表的一项比较真实针灸和假针灸的试验发现，前者的神经病理性疼痛减少了一半，这个结果令人印象深刻。