Does glycemic control affect outcome of COVID‐19?

Diabetes is associated with adverse outcome of coronavirus disease 2019 (COVID‐19). In a whole‐population study of a UK data set of >60 million persons, approximately 5% had diabetes, but of 23 698 in‐hospital COVID‐19‐related deaths in this population, 33.2% were of people with diabetes, with mortality rates among persons those with type 1 and type 2 diabetes approximately 5‐fold and 10‐fold greater than that in the nondiabetic population; with adjustment for age and other risk factors, type 1 and type 2 diabetes were associated with near tripling and doubling of mortality rates, respectively. ¹

We define diabetes by the presence of hyperglycemia, but we also know that, even among persons not known to have diabetes, the development of hyperglycemia in the setting of infection, trauma, surgery, or a cardiovascular event is associated with adverse outcome. It is then logical to ask whether the adverse outcome of COVID‐19 among persons with diabetes is related to the degree of hyperglycemia and, if so, whether this represents a causal relationship or rather shows that the severity of illness in a given individual with diabetes leads to both higher blood glucose concentration and greater likelihood of mortality. So far, reports from different data sets are somewhat contradictory.

The same whole‐population UK data set reported that, compared with those having glycated hemoglobin (HbA1c) 48‐53 mmol/mol (6·5‐7·0%), COVID‐19‐related mortality with HbA1c of 86 mmol/mol (10·0%) or more was 2.23 fold greater among persons with type 1 diabetes, and among those with type 2 diabetes, mortality with HbA1c 59‐74 (7.6%‐8.9%), 75‐85 (9.0‐9.9), and ≥86 (10.0%) was increased 1.22, 1.36, and 1.61‐fold, respectively. ² A different UK data set of >17 million adult National Health Service patients gave an adjusted mortality hazard ratios of 1.31 and 1.95 among those with diabetes having HbA1c <7.5% and ≥7.5%, respectively. ³

Specific hospitalized patient data sets, however, do not show that prior glycemic control, as reflected in HbA1c, are associated with different likelihood of adverse outcome among persons with diabetes. In an Iranian hospital, 24 persons with HbA1c 6.5‐7 had a similar mortality rate (21%) to that of 93 persons (23%) with HbA1c 8%‐11.2%. ⁴ In analysis of 1317 persons with diabetes hospitalized for COVID‐19 in France, the outcome of tracheal intubation or death by day 7 was not associated with HbA1c level on admission, although there was a linear correlation with admission plasma glucose, as well as with body mass index. ⁵ Another study, of 1279 persons with diabetes hospitalized for COVID‐19 in New York, showed no association of mortality with HbA1c on admission, whereas obesity, as well as increasing age and male sex and increasing complexity of outpatient diabetes treatment, were again associated with greater likelihood of adverse outcome. ⁶

What can we conclude from the discrepancy between the population and hospital data sets? Conceptually, several sets of mediators of adverse COVID‐19 outcome may exist among persons with diabetes (Box 1). There may be a direct risk associated with hyperglycemia. Hyperglycemia may, however, be a marker of stress of illness. The risk seen among persons with diabetes may be because of insulin resistance, inflammation, hypercoagulation, or underlying obesity. And diabetes is associated with cardiovascular disease, chronic kidney disease, dyslipidemia, hypertension, and a variety of other conditions, which may mediate adverse outcome.

BOX 1. Potential mediators of adverse COVID‐19 outcome in diabetes.

Hypothesis	Implications
Direct increased risk from hyperglycemia	Goal: normalize glycemia (subhypothesis: adverse outcome proportionate to degree of hyperglycemia)
Hyperglycemia as marker of metabolic stress	Severity of illness rather than glycemia per se causing adverse outcome
Risk because of obesity, insulin resistance, inflammation, or hypercoagulation	Glycemia per se may not directly lead to adverse outcome
Diabetes as marker of associated illnesses	Cardiovascular disease, chronic kidney disease, dyslipidemia, hypertension, etc. rather than glycemia per se as actual mediators

Certainly, glycemic control matters, and we have learned a great deal about practical approaches to optimizing glycemia during hospitalization with COVID‐19. ⁷ Whether or not glycemia is the direct mediator, the disturbing association of diabetes with adverse outcome of COVID‐19 is of great concern and must inform our ongoing approach to the care of our patients.

糖尿病与新冠肺炎(COVID‐19)的不良预后有关。在一项对英国6000万人的全人群研究中，大约5%的人患有糖尿病，但在23698例与新冠肺炎有关的住院死亡中，33.2%是糖尿病患者，1型和2型糖尿病患者的死亡率分别是非糖尿病人群的5倍和10倍；在调整了年龄和其他危险因素后，1型和2型糖尿病的死亡率分别增加了近3倍和2倍。

虽然我们以出现高血糖来定义糖尿病，但我们也知道，即使在未患有糖尿病的人中，在感染、创伤、手术或心血管事件中发生的高血糖也与不良预后有关。那么顺理成章的问题是，新冠肺炎在糖尿病患者中的不良预后是否与高血糖的程度有关，如果是，那么这是否是因果关系，或者更确切地说，这是否表明特定糖尿病患者的疾病严重程度会导致更高的血糖和死亡可能性。然而到目前为止，来自不同研究的报告还是存在某些矛盾。

来自全人群英国数据集的报告指出，与糖化血红蛋白(HbA1c)48‐53 mmol/mol(6.5‐7.0%)相比，1型糖尿病患者中HbA1c≥86 mmol/mol(10.0%)的新冠肺炎相关死亡率高出2.23倍，而在2型糖尿病患者中，HbA1c 59‐74(7.6%‐8.9%)、75‐85(9.0‐9.9%)和≥86(10.0%)的死亡率分别增加1.22、1.36以及1.61倍；另一项英国国家医疗服务体系(NHS)的1700万成人患者的数据集显示，在HbA1c<7.5%和≥7.5%的糖尿病患者中，校正后的死亡风险比分别为1.31和1.95。

然而，住院患者的特定数据集并未显示HbA1c反映的既往血糖控制水平与糖尿病患者不良预后的相关性。在一家伊朗医院，24名HbA1c 6.5%‐7%的患者死亡率(21%)与93名HbA1c 8%‐11.2%的患者(23%)死亡率相似。在法国因新冠肺炎住院的1317名糖尿病患者中，第7天气管插管或死亡的结局与入院时的HbA1c水平无关，但与入院时的血糖水平以及体重指数呈线性相关。在纽约因新冠肺炎住院的1279名糖尿病患者中，死亡率与HbA1c无关，而肥胖、高龄、男性性别以及门诊糖尿病治疗的复杂性则与不良结局相关。

我们可以从人口和医院数据集之间的差异中得出什么结论?从概念上讲，糖尿病患者中可能存在几组与新冠肺炎不良预后相关的因素，可能有与高血糖相关的直接风险。然而，高血糖可能是疾病压力的一个标志。糖尿病患者的风险可能是胰岛素抵抗、炎症、高凝状态或潜在的肥胖。糖尿病与心血管疾病、慢性肾脏疾病、血脂异常、高血压和其他各种疾病有关，这些疾病可能会导致不良预后。

当然，血糖控制很重要，我们已经学到了很多关于在感染新冠后住院期间优化血糖的实用方法。无论血糖是否是直接调节因素，糖尿病与新冠肺炎不良预后的关联都是令人非常担忧的，因此必须了解这些信息，照护好患者。