We read with interest the recent commentary published in the journal wherein Maddaloni and Buzzetti¹ describe the potential increase in the risk of new onset diabetes secondary to coronavirus disease 2019 (COVID-19) pandemic. In Singapore, as of 9 November 2020, there are a total of 58,064 confirmed COVID-19 cases, with 57,981 recoveries and 28 deaths (https://www.moh.gov.sg/covid-19/situation-report). In late March and April, clusters of COVID-19 infection were detected in migrant workers residing at dormitories which contributed to an overwhelming proportion of cases in the country. We describe the clinical presentations of three South Asian migrant workers who presented with diabetic ketoacidosis (DKA) and marked insulin resistance and characteristics reminiscent of type 1B diabetes with mild COVID-19 infection.

These three young physically active workers had a mean age 35 (8.1) years and were lean with a mean body mass index (BMI) of 20.88 (SD: 5.22) kg/m². None of the patients had a personal history or family history of diabetes mellitus (DM). They presented with mild symptoms of upper respiratory tract infection and/or loss of smell with onset 7–10 days before presentation. The nasopharyngeal swab test (polymerase chain reaction) for severe acute respiratory syndrome coronavirus 2 was positive. The initial venous plasma glucose was more than 20 mmol/L and glycated haemoglobin (HbA1c) was more than 11% in all subjects. Laboratory evaluation confirmed diagnosis of moderately severe DKA (Kitabchi's classification).² They all had ketosis with beta hydroxybutyrate (BHB) concentrations between 5 and 7 mmol/L and mild acidosis (arterial pH 7.2–7.3; bicarbonate 9–12 mmol/L). Paradoxically, this was not accompanied by a systemic inflammatory response as C-reactive protein and total white cell counts were not significantly increased. Ketoacidosis resolved within 2–3 days of institution of treatment as per DKA guidelines constituting of intravenous insulin infusion, rehydration with intravenous fluids and correction of electrolytes (Table 1). COVID-19 infection remained mild in all patients and none of them had hypoxia or clinical pneumonia. They recovered within 2 weeks of admission for COVID-19. The insulin requirements on discharge was more than 0.6 units/kg for all patients. Two months after discharge, the fasting serum C-peptide concentrations were low for all patients. Islet cell antibodies, namely islet cell antibody, glutamic acid decarboxylase autoantibodies (GADA), insulinoma-associated antigens 2 autoantibody (IA-2ßA) and zinc transporter autoantibody were measured by radioligand assay according to the Islet Autoantibody Standardization Program. The sensitivity and specificity of GAD antibody assay were 76.0% and 87.8%, IA2 antibody assays were 76.0% and 94.4%, respectively, as evaluated in the Fourth Diabetes IASP 2015 (laboratory ID 1501).³ These islet cell autoantibodies were all absent. Thyroid function tests were normal in all patients. Hence, they could all be classified as nonautoimmune type 1B diabetes (Table 1).⁴ Upon follow up, all of the patients still require insulin although there is a slight reduction in the dose at approximately 0.5 units/kg body weight.

We searched the literature for similar cases of new onset diabetes presenting with DKA exacerbated by COVID-19 infections. Hollstein et al.⁵ reported a 19 years old Caucasian man who presented 4 weeks after COVID-19 infection (with mild symptoms) with DKA. The immunoglobulin G antibodies for SARS-CoV was positive confirming past infection. Similar to our case, islet cell autoantibody was negative.⁵ An increase in the incidence of new Type 1 diabetes (compared to usual incidence) has been reported between March and June 2020, at five paediatric inpatient units from four National Health Service Trusts in London, UK but specific cases of nonautoimmune type 1B diabetes are not known.⁶

All these subjects have mild COVID19 disease but developed moderately severe DKA and required high doses of insulin for glycaemic control. They had low C-peptide levels and no evidence of islet cell directed autoimmunity. This suggests the presence of type 1B insulin-dependent diabetes (nonautoimmune type 1 diabetes) in direct association with the coronavirus infection.⁷ Type 1B diabetes or ketosis-prone DM is a heterogeneous syndrome reported in African Americans, Hispanic descendants and in Asians. They may harbour a combination of characteristics of type 1 or type 2 diabetes. It remains to be seen whether the reported patients will require long-term insulin or will have periods with no insulin requirements. The relatively high insulin requirement suggests impairment of insulin signalling due to concomitant insulin resistance. It is known that the SARS-CoV-2 virus utilises the angiotensin-converting enzyme 2 (ACE2) which is highly expressed in the pancreas⁸ and may protect the function of insulin-producing beta cells by improving the function of islet microvascular endothelial cells. In the skeletal muscle, as a major target of insulin action, the ACE-2-Ang(1-7)-Mas axis has been described to decrease insulin resistance.⁸ Hence the use of these receptors and subsequent downregulation of these pathways may partially explain this presentation.

A comparison of non immune-mediated versus immune-mediated type 1 diabetes has shown that these patients tend to be older than immune-mediated diabetes, but they have a higher cardiovascular risk resembling the phenotype of type 2 diabetes.⁹ Hence, it would be important to follow up these patients for development of vascular complications or sequelae especially since COVID-19 is known to exacerbate the thrombotic and inflammatory pathways in the endothelium.¹⁰

In summary, these three healthy and fit, lean South Asians presented with type 1B insulin-dependent diabetes. This clinical presentation adds credence to the hypothesis that the SARS-CoV-2 virus has the potential to directly affect pancreatic function and peripheral insulin resistance state. The mechanisms involved needs to be further elucidated.¹

我们感兴趣地阅读了最近发表在杂志上的评论，其中 Maddaloni 和 Buzzetti ¹描述继发于 2019 年冠状病毒病 (COVID-19) 大流行的新发糖尿病风险的潜在增加。截至 2020 年 11 月 9 日，新加坡共有 58,064 例确诊的 COVID-19 病例，57,981 人康复，28 人死亡（https://www.moh.gov.sg/covid-19/situation-report）。在 3 月下旬和 4 月，在居住在宿舍的农民工中发现了 COVID-19 感染群，这导致该国病例的比例居高不下。我们描述了三名南亚移民工人的临床表现，他们出现糖尿病酮症酸中毒 (DKA)，并具有明显的胰岛素抵抗和特征，使人联想到 1B 型糖尿病伴轻度 COVID-19 感染。

这三名年轻的体力活动工作者的平均年龄为 35 (8.1) 岁，平均体重指数 (BMI) 为 20.88 (SD: 5.22) kg/m ²。没有患者有糖尿病（DM）的个人病史或家族史。他们在就诊前 7-10 天出现上呼吸道感染和/或嗅觉丧失的轻微症状。严重急性呼吸综合征冠状病毒 2 的鼻咽拭子试验（聚合酶链反应）呈阳性。所有受试者的初始静脉血浆葡萄糖超过 20 mmol/L，糖化血红蛋白 (HbA1c) 超过 11%。实验室评估证实了中重度 DKA 的诊断（Kitabchi 分类）。²他们都患有酮症，β-羟基丁酸 (BHB) 浓度在 5 至 7 mmol/L 和轻度酸中毒（动脉 pH 7.2-7.3；碳酸氢盐 9-12 mmol/L）。矛盾的是，这并没有伴随全身炎症反应，因为 C 反应蛋白和总白细胞计数没有显着增加。根据 DKA 指南，包括静脉注射胰岛素、静脉补液和纠正电解质，酮症酸中毒在 2-3 天内得到缓解（表 1）。所有患者的 COVID-19 感染都很轻微，没有人出现缺氧或临床肺炎。他们在 COVID-19 入院后 2 周内康复。所有患者出院时的胰岛素需求量均超过 0.6 单位/公斤。出院两个月后，所有患者的空腹血清 C 肽浓度均较低。胰岛细胞抗体，即胰岛细胞抗体、谷氨酸脱羧酶自身抗体 (GADA)、胰岛素瘤相关抗原 2 自身抗体 (IA-2ßA) 和锌转运蛋白自身抗体根据胰岛自身抗体标准化计划通过放射性配体测定进行测量。GAD 抗体检测的敏感性和特异性分别为 76.0% 和 87.8%，IA2 抗体检测分别为 76.0% 和 94.4%，如第四届糖尿病 IASP 2015（实验室 ID 1501）所评估。³这些胰岛细胞自身抗体均不存在。所有患者的甲状腺功能检查均正常。因此，它们都可以归类为非自身免疫性 1B 型糖尿病（表 1）。⁴随访时，所有患者仍需要胰岛素，尽管剂量略有减少，约为 0.5 单位/kg 体重。

我们在文献中搜索了类似的新发糖尿病病例，这些病例表现为因 COVID-19 感染而加重的 DKA。霍尔斯坦等人。⁵报告了一名 19 岁的白人男子，他在 COVID-19 感染（症状轻微）后 4 周出现 DKA。SARS-CoV 的免疫球蛋白 G 抗体呈阳性，证实了过去的感染。与我们的情况相似，胰岛细胞自身抗体是阴性的。⁵据报道，2020 年 3 月至 2020 年 6 月期间，英国伦敦四个国家卫生服务信托基金的五个儿科住院病房的新 1 型糖尿病发病率（与正常发病率相比）有所增加，但非自身免疫性 1B 型糖尿病的特定病例是未知。⁶

所有这些受试者都患有轻度 COVID19 疾病，但发展为中度严重的 DKA，需要高剂量的胰岛素来控制血糖。他们的 C 肽水平低，没有胰岛细胞定向自身免疫的证据。这表明存在与冠状病毒感染直接相关的 1B 型胰岛素依赖型糖尿病（非自身免疫性 1 型糖尿病）。⁷1B 型糖尿病或酮症倾向 DM 是一种在非裔美国人、西班牙裔后裔和亚洲人中报告的异质综合征。他们可能同时具有 1 型或 2 型糖尿病的特征。报告的患者是否需要长期使用胰岛素或有一段时间不需要胰岛素还有待观察。相对较高的胰岛素需求表明由于伴随的胰岛素抵抗导致胰岛素信号传导受损。众所周知，SARS-CoV-2 病毒利用在胰腺中高度表达的血管紧张素转换酶 2 (ACE2) ⁸并可能通过改善胰岛微血管内皮细胞的功能来保护产生胰岛素的β细胞的功能。在骨骼肌中，作为胰岛素作用的主要目标，ACE-2-Ang(1-7)-Mas 轴已被描述为降低胰岛素抵抗。⁸因此，这些受体的使用和随后这些途径的下调可能部分解释了这种表现。

非免疫介导与免疫介导的 1 型糖尿病的比较表明，这些患者往往比免疫介导的糖尿病患者年龄更大，但他们的心血管风险更高，类似于 2 型糖尿病的表型。⁹因此，对这些患者进行血管并发症或后遗症的随访非常重要，尤其是因为已知 COVID-19 会加剧内皮中的血栓形成和炎症通路。¹⁰

总之，这三名健康、体格健壮的瘦南亚人患有 1B 型胰岛素依赖型糖尿病。这种临床表现增加了对 SARS-CoV-2 病毒有可能直接影响胰腺功能和外周胰岛素抵抗状态的假设的可信度。所涉及的机制需要进一步阐明。¹