key: cord-0850941-ylms8pb0
authors: Lazzeri, Chiara; Bonizzoli, Manuela; Batacchi, Stafano; Di Valvasone, Simona; Chiostri, Marco; Peris, Adriano
title: THE PROGNOSTIC ROLE OF HYPERGLYCEMIA AND GLUCOSE VARIABILITY IN COVID-RELATED ACUTE RESPIRATORY DISTRESS SYNDROME
date: 2021-04-01
journal: Diabetes Res Clin Pract
DOI: 10.1016/j.diabres.2021.108789
sha: 3861acfaec7ef317639b1946547885a5fda03786
doc_id: 850941
cord_uid: ylms8pb0

Aims Due to heterogeneity on the prognostic role of glucose values and glucose variability in Novel Coronavirus (COVID) disease, we aimed at assessing the prognostic role for Intensive Care Unit (ICU) death of admission hyperglycaemia, peak glycemia and glucose variability in critically ill COVID patients: Methods 83 patients consecutively admitted for COVID-related Acute Respiratory Distress Syndrome (ARDS) from from 1st March to 1st October 2020. Results Non survivors were older, with more comorbidities and a more severe disease. Corticosteroids were used in the majority of patients (54/83, 65%) with no difference between survivors and non survivors. Mean blood glucose values, (during the first 24 and 48 hours, respectively), were comparable between the two subgroups, as well as SD 24 and CV 24. During the first 48 hours, survivors showed significantly lower values of SD 48 (p<0.001) and CV 48, respectively (p<0.001) than non survivors. Conclusions in consecutive COVID-related ARDS patients admitted to ICU hyperglycemia (>180 mg/dl) is more common in non survivors who also showed a significantly higher glucose variability in the first 48 hours since ICU admission. Our findings point to the clinical significance of in-ICU glucose control in severe COVID patients.

Hyperglycaemia (stress-related hyperglycemia) is quite common in critically ill patients, and, though in the lack of an universal threshold for stress-related hyperglycamia, it was associated with increased morbidity and mortality in critical illness [1] [2] [3] . Also glycemic variability was recognized as a strong independent predictor of mortality among critically-ill patients [4, 5] .

To date, evidence on the prognostic role of hyperglycemia in COVID disease is scarce and heterogenous mainly due to differences in clinical characteristics of study populations (ie. with or without previously known diabetes, and different degrees of disease severity).

The present investigation was aimed at assessing the prognostic role for ICU death of admission hyperglycaemia, peak glycemia and glucose variability (as indicated of the standard deviation of mean glucose levels and the coefficient of variation of glucose) in 83 patients consecutively admitted for COVID-related Acute Respiratory Distress Syndrome (ARDS) in our ICU (which is an ECMO referral center) from 1 st March to 1 st October 2020.

In this case series study we enrolled all patients with COVID-19 ARDS consecutively admitted to our ICU (which is an ECMO referral center) from 1 st March to 1 st October 2020. The study protocol was approved by our Ethical Committee (n.17024, approved on March 31 th 2020). ARDS was defined according to the Berlin definition [6] .

For each patient the following clinical variables were recorded in a dedicated database: age, gender, body mass index (BMI) and risk factors (previously known diabetes, hypertension, history of heart disease). Charlson comorbidity index was calculated based on medical history [7] . The simplified acute physiology score (SAPS) II was also calculated [8] .

On ICU admission we measured: C-reactive Protein (CRP, mg/dl) creatinine (mg/dl), Lactose dehydrogenase (LDH, UI/L), alanine transaminase (ALT, UI/L), Ddimer (ng/ml)and interleukin 6 (IL-6, pg/ml).

A predefined insulin protocol was used in all patients for management of hyperglycemia and intensive insulin therapy was administered in patients with significant hyperglycemia (that is plasma glucose >150 g/l) [1] [2] [3] 9] .

According to our protocol [2, 9] , glucose values were measured four times a day, and the peak glucose was determined among all values measured during ICU stay [10] [11] [12] . Glucose values for the first 24 and 48 hours were recorded to calculate the following parameters: a) Mean blood glucose level as arithmetic mean of all recorded glucose values for each patient for the first 24 hours (Mean BG 24) and for the first 48 hours (Mean BG 48); b) Standard deviation (SD) of mean glucose levels for the first 24 hours (SD 24) and for the first 48 hours (SD 48); c) Coefficient of variation (CV) of glucose (derived as a percentage of SD to mean blood glucose) for the first 24 hours (CV 24 ) and for the first 48 hours (CV48 and SD48 (CV 48) [2, 4, 5, 13] .

Data were prospectively recorded and retrospectively analyzed.

The primary endpoint for the analysis was in-ICU mortality (defined as death before hospital discharge). 

Our population comprises 83 consecutive patients with COVID-related ARDS admitted to our ICU ( Table 1 ). The study population included mainly males (83%), and hypertension was the commonest risk factor, being detectable in 78% of cases.

Previously known diabetes was observed in 33 patients (40%). Most patients (62%) were mechanically ventilated. In our series ICU mortality was 36% (30/83).

In the comparison between survivors and non survivors (Table 1) , non survivors were older (p<0.001), with more comorbidities, as indicated by a higher Charlson's index (p<0.001) and with a more severe disease, inferred by a higher SAPS II index (p<0.001). Higher values of creatinine were observed in no survivors (p=0.032).

Corticosteroids were used in the majority of patients (54/83, 65%) with no difference between survivors and non survivors (survivors: 33; non survivors: 21, p= 0.477, chi square test). Table 2 shows glucose values and glucose variability parameters (both at 24 and 48 hours since ICU admission) in the overall population and in the comparison between survivors and non suvivor patients. No difference was observed in the two subgroups in admission glucose values but the frequency of admission glycemia > 180 mg/dl was significantly higher in non survivors (p=0.036). Peak and nadir glycemia during ICU stay were comparable between the two subgroups, but the frequency of hypoglycemia (< 70 mg/dl) was higher in non survivors (p=0.042). Mean blood glucose values, both during the first 24 and 48 hours, respectively, were comparable between survivors and non survivors, as well as SD 24 and CV 24. During the first 48 hours since ICU admission, survivors showed significantly lower values of SD 48 (p<0.001) and CV 48, respectively (p<0.001).

The main findings of the present investigation, performed in consecutive COVID-related ARDS patients admitted to ICU, are as follows: a) hyperglycemia (>180 mg/dl) is more common in non survivors; b) glucose variability in the first 48 hours since ICU admission is significantly higher in non survivors. Overall, our findings point to the clinical significance of glucose control in severe COVID patients during ICU stay.

To date, evidence on the prognostic role of hyperglycemia in COVID disease is scarce and heterogenous mainly due to differences in clinical characteristics of study populations (ie. patients with or without previously known diabetes, and different degrees of disease severity).

In non critically ill patients, elevated glucose values were associated with a worse prognosis [14] [15] [16] [17] [18] [19] [20] [21] [22] , but heterogeneity in study design can be detected in these investigations concerning study population (diabetics versus non diabetics) and blood glucose measurements (admission, fasting, in hospital glucose values). In a two-center retrospective investigation [17] (605 COVID patents without previous diagnosis of diabetes) fasting blood glucose (FBG ≥ 7.0 mmol/l) at admission was an independent predictor for 28-day mortality . However, no data were provided on the percentage of patients with severe COVID disease (admitted to ICU) nor on the standard protocol of in-hospital glucose management. The association between hyperglycemia and increased risk of death was confirmed in further studies performed in non severe COVID disease [18, 19] and in non diabetics [21] , but evidence is overall not univocal. To date, glycemic variability was specifically investigated in COVID disease only by Zhu et al [15] who, in a retrospective analysis, documented that in patients with pre-existing type 2 diabetes, glycaemic variability during hospitalization within 3.9 -10.0 mmol/l (that is well-controlled glucose) was associated with markedly lower mortality compared with patients with poorly controlled blood glucose (glycaemic variability more than mmol/l) [15] . However, patients aged >75 years, and those with acute lethal organ injury or acute decompensated or end stage organ dysfunction were excluded. Recently, Shen et al [26] observed, by means of continuous glucose monitoring, that patients of diabetes and COVID disease (35 patients) had an increase risk of outcomes with glucose values > 160 mg/dl and < 70 mg/dl and a high coefficient of variation.

In this context, this is the first investigation assessing glucose variability in consecutive COVID patients with severe disease, admitted to ICU. We observed that 48h glucose variability (as indicated by SD48 and CV48) was higher in non survivors compared to survivors. This finding may underscore the prognostic role of glucose management in these critically ill patients. 

A systematic scoping review on the consequences of stress-related hyperglycaemia

The prognostic role of peak glycemia and glucose variability in trauma: a single-center investigation

Glucose control in the ICU

Blood glucose amplitude in critically ill patients

Glucose variability measures and their effect on mortality: a systematic review

Acute respiratory distress syndrome: the Berlin Definition

A new method of classifying prognostic comorbidity in longitudinal studies: development and validation

A new Simplified acute Physiology score (saPs ii) based on a european/North american multicenter study

Early 10 predictor? A prospective, monocentric pilot study

The prognostic role of in-hospital peak glycemia in stemi patients with and without diabetes

In-hospital peak glycemia and prognosis in STEMI patients without previously known diabetes

In-hospital peak glycemia in non-diabetic patients with heart failure complicating acute coronary syndrome

Early glucose variability in cardiogenic shock following acute myocardial infarction: a pilot study

Clinical Study of Hyperglycemia and SARS-CoV-2 Infection in Intensive Care Unit Patients In vivo

Association of blood glucose control and outcomes in patients with COVID-19 and pre-existing type 13

Dysglycaemia in the critically ill and the interaction of acute and chronic glycaemia with mortality

NIV: non invasive ventilation, P/F: PO2/FiO2 ration, SAPS II: simplified acute physiologyscore; LOS: lenght of stay, LDH: lactic dehydrogenase; CRP: C Reactive protein, IL-6: interleukin 6, IQR: interquartile rante, KW: Kruskal-Wallis test