key: cord-0946672-pklxh60l
authors: Prattichizzo, Francesco; de Candia, Paola; Nicolucci, Antonio; Ceriello, Antonio
title: Elevated HbA1c levels in pre‐Covid‐19 infection increases the risk of mortality: A sistematic review and meta‐analysis
date: 2021-05-28
journal: Diabetes Metab Res Rev
DOI: 10.1002/dmrr.3476
sha: 4dce62a712492bf2c371ab3030a77b89d5e401ec
doc_id: 946672
cord_uid: pklxh60l

AIMS: Diabetes is emerging as a risk factor for coronavirus disease (COVID)‐19 prognosis. However, contradictory findings have been reported regarding the impact of glycaemic control on COVID‐19 outcome. The aim of this meta‐analysis was to explore the impact of hospital pre‐admission or at‐admission values of HbA1c on COVID‐19 mortality or worsening in patients with diabetes. MATERIALS AND METHODS: We searched PubMed, Embase and Scopus up to 30th December 2020. Eligibility criteria for study selection were the following: (1)enrolling patients with any form of diabetes mellitus and hospitalized for COVID‐19 and (2) reporting data regarding HbA1c values before infection or at hospital admission in relation to COVID‐19 mortality or worsening. Descriptive statistics, HbA1c values, odds ratios (ORs) and hazard ratios were extracted from seven observational studies and generic inverse variance (random effects) of OR was used to estimate the effect of HbA1c on COVID‐19 outcome. RESULTS: HbA1c was linearly associated with an increased COVID‐19 mortality or worsening when considered as a continuous variable (OR 1.01 [1.01, 1.01]; p < 0.00001). Similarly, when analysing studies providing the number of events according to the degree of glycaemic control among various strata, a significantly increased risk was observed with poor glycaemic control (OR 1.15 [1.11, 1.19]; p < 0.00001), a result corroborated by sensitivity analysis. CONCLUSIONS: Notwithstanding the large heterogeneity in study design and patients' characteristics in the few available studies, data suggest that patients with diabetes and poor glycaemic control before infection might have an increased risk of COVID‐19 related mortality.

Diabetes mellitus (DM) is emerging as a critical risk factor for coronavirus diseases (COVID)-19 poor prognosis, with a recent metaanalysis reporting that COVID-19 patients with pre-existing DM have a threefold increased risk of in-hospital mortality. 1,2 A number of hypotheses has been proposed to explain the observed increased risk among patients with DM, and multiple variables explored as intermediate risk factors. 3, 4 Among others, blood glucose levels are emerging as a critical prognostic factor for COVID-19 mortality in both patients with and without DM. [4] [5] [6] On the other hand, conflicting data have been reported regarding hospital pre-admission or atadmission assessment of glycaemic control in relation to related mortality in patients with DM. 7, 8 The objective of this study was to explore whether glycaemic control, as measured by HbA1c, is a relevant prognostic factor for acute COVID-19 mortality or worsening of symptoms in patients with DM.

We searched for studies through PubMed, Embase and Scopus up to 30th December 2020, collecting only articles in English. The strings used for the PubMed search can be found in the Data S1. Twentyfour abstracts were identified. Two investigators (F.P. and P.d.C.)

independently reviewed the selected abstracts to determine the eligibility of the studies for the meta-analysis according to two inclusion criteria: (1) enrolling patients with any form of pre-existing DM and hospitalized for COVID-19 (laboratory confirmed or clinically assigned) and (2) reporting data regarding HbA1c values at hospital admission or before infection in relation to COVID-19 mortality or a composite outcome including mortality with any follow-up duration, including studies not having this outcome as the primary endpoint. Exclusion criteria were the following: (1)-populations composed of COVID-19 patients without pre-existing DM and (2)-manuscripts not reporting HbA1c among the variables analysed. The primary outcome of our analysis was mortality or the composite of mortality and disease worsening/progression, with no restriction to the definition of worsening. No secondary outcome was collected. The protocol was registered in OSF and it is accessible at https://osf.io/8h9yn. The MOOSE checklist 9 is available as Data S1.

Two authors (F.P. and P.d.C.) used a standardized collection form to extract summary estimates data of selected studies. Included information were study design, sample size, primary outcome, sex, age, diabetes duration, BMI, HbA1c, reported hazard ratios (HRs) or odds ratios (ORs) for HbA1c, and the relative statistic approach, including the adjustments applied. Data were checked for accuracy by two additional investigators (A.C. and A.N.). F.P. and P.d.C. independently used the Newcastle-Ottawa Scale (NOS) for non-randomized cohort studies 10 to perform quality assessment of included manuscripts.

We used the generic inverse variance method to estimate the effect of HbA1c on COVID-19 mortality starting from collected OR with the relative confidence intervals (CI)s. As the statistical methods were different in the collected studies, we separately analysed studies analysing HbA1c measurements as a continuous variable and studies instead providing risk estimates according to the degree of glycaemic control, categorizing HbA1c as a dichotomous variable. For those studies reporting HR for multiple HbA1c strata, 8, 11 we calculated the relative crude OR with 95% CIs by extracting the number of events and the number of patients for two groups, split according to an HbA1c value of either < or >7.5% (good vs. poor glycaemic control, respectively). This cut-off was selected in order to use all the collected studies and minimize the risk of bias since one study only allowed this extraction, 11 albeit current guidelines recommend an HbA1c target of 7% for most patients with DM. 12 Statistical heterogeneity between studies was evaluated by I 2 statistic and the significance for heterogeneity was set at I 2 >50% or p < 0.1, using the fixed effect model to estimate summary below this limit and the random effects model above the same threshold. For the sensitivity analysis, we performed an alternative calculation for studies reporting HR for multiple HbA1c strata, using HR as OR (given the short time of follow-ups, i.e., COVID-19 mortality or worsening is observed in 7-30 days) and pooling them through the generic inverse variance method (fixed effect) to obtain one OR value against the reference group. 8, 13 All analyses were performed using review manager 5.4 (Cochrane Collaboration). 14 3 | RESULTS

Inclusion flow of studies is presented in Figure S1 . Of the 26 abstracts identified, 16 manuscripts reported only data regarding blood glucose levels or range in relation to the observed outcome and not HbA1c values, 4,15-30 one study has only intubation as reported outcome, 31 another one included patients without DM in the group with good glycaemic control, 32 while the remaining one was focused on the comparison between pre-existing and new-onset diabetes. 33 

The characteristics of the included seven manuscripts 7, 8, 11, 13, [34] [35] [36] are presented in Table 1 which provided separate data for type 1 DM (T1DM) and DM overall in two different manuscripts, 13, 34 while another large study provided separate data for T1DM and T2DM in the same manuscript, 8 thus yielding an overall of eight study groups for the meta-analysis. The quality assessment of included studies is reported in Figure S2 .

Overall, the analysis involved 4,985,063 patients, 298,850 with T1DM and 1,533,579 with a non-specified form of DM. As evidenced in Table 1 , there was a large heterogeneity in terms of study design, patients' characteristics and the approach adopted to both report descriptive statistics and calculate OR for the selected clinical variables. Thus, we analysed data treating HbA1c as a continuous variable separated from those allowing to treat glycaemic control as a dichotomous variable. For the same reasons, it was not possible to calculate the mean values of the extracted variables. As two manuscripts might have enrolled the same populations 8,13 albeit using different databases, we considered these studies one at a time.

When considering HbA1c as a continuous variable, the results of the meta-analysis showed that higher HbA1c values were associated with an increased COVID-19 related mortality or worsening (OR p < 0.00001; I 2 = 57%; p = 0.13; Figure 2B ). Then, we performed an additional sensitivity analysis, by calculating OR for two studies 8, 13 with a different approach, which allowed to include additional mortality data from the Coronado study, 13 p < 0.0003; I 2 = 6%; p = 0.35; Figure 2D ).

Finally, when including only studies measuring HbA1c at hospital admission, we did not obtain a significant result ( Figure S3 ).

COVID-19 pandemic is affecting an ever-increasing number of people worldwide. Patients with DM are among those mostly suffering the consequences of the infection, with an increased risk of mortality compared to the general population. 2 +By pooling adjusted OR through generic inverse variance, fixed effect, against the reference group.

prognosis, 5 the prognostic value of glycaemic control is uncertain. In this meta-analysis, we showed that HbA1c values measured prior to or at-hospital admission are linearly associated with an increased risk of COVID-19 mortality or worsening. In addition, comparing patients with different degree of glycaemic control suggested that subjects with a poor glycaemic control might have an increased risk of COVID-19 related mortality.

These findings must be interpreted with caution for a variety of reasons. There was a large heterogeneity in study design, patients' characteristics and reporting of descriptive statistics. As a result, we had to recalculate OR from large studies 8, 11 by selecting a cut-off for HbA1c (7,5%) that allowed the inclusion of all studies. We are not able to exclude the possibility that setting a lower cut-off would have yielded different results, but extracted data did not allow a different analysis. In addition, the remaining study, 35 reported data only setting a cut-off for glycaemic control at HbA1c < or >8%. However, the sensitivity analysis, performed by calculating OR for these studies with an alternative approach taking advantage of adjusted HR, provided similar results. On the other hand, the degree of glycaemic control for the reference group was not homogenous even in this case. 8, 13 Furthermore, the two larger studies, which clearly influenced the results, derived both from English electronic health records. 8, 11 Albeit their relative source databases were different, 8 23, 39 In addition, the analyses did not take in account the different classes of glucose-lowering agents used before infection, an aspect that is also emerging as a possible determinant of COVID-19 prognosis. 23, 28 An additional factor that might have influenced the results is the length of follow-up. Indeed, the only prospective study 13 has only a small, albeit significant, linear association with COVID-19

prognosis when performing multiple adjustments. As also suggested by others, 7 considering that HbA1c has been associated with a higher basal state of low-grade inflammation 42 and a dysregulated immune cell function, 43, 44 it is reasonable to expect a higher risk of complicated prognosis in DM patients with a poor glycaemic control.

However, low-grade inflammation and immune cell function in turn depend on a plethora of variables, including all those mentioned above. 45 additional, prospective studies are needed to fully establish a cor-

Diabetes is a risk factor for the progression and prognosis of COVID-19

Diabetes as a risk factor for greater COVID-19 severity and in-hospital death: a metaanalysis of observational studies

Severe obesity, increasing age and male sex are independently associated with worse inhospital outcomes, and higher in-hospital mortality

Newly-diagnosed diabetes and admission hyperglycemia predict COVID-19 severity by aggravating respiratory deterioration

Effects of hyperglycaemia on complications of COVID -19: a meta-analysis of observational studies

Why is hyperglycaemia worsening COVID -19 and its prognosis?

Preadmission diabetes-specific risk factors for mortality in hospitalized patients with diabetes and coronavirus disease 2019. Dia Care

Risk factors for COVID-19-related mortality in people with type 1 and type 2 diabetes in England: a population-based cohort study

Meta-analysis of observational studies in EpidemiologyA proposal for reporting

Factors associated with COVID-19-related death using OpenSAFELY

Glycemic targets: standards of medical care in diabetes-2020

Phenotypic characteristics and prognosis of inpatients with COVID-19 and diabetes: the CORO-NADO study

Hyperglycemia at hospital admission is associated with severity of the prognosis in patients hospitalized for COVID-19: the Pisa COVID-19 study. Dia Care

Elevation of blood glucose level predicts worse outcomes in hospitalized patients with COVID-19: a retrospective cohort study

Glycemic characteristics and clinical outcomes of COVID-19 patients hospitalized in the United States

Hyperglycaemia on admission to hospital and COVID-19

Fasting blood glucose at admission is an independent predictor for 28-day mortality in patients with COVID-19 without previous diagnosis of diabetes: a multi-centre retrospective study

The impact of type 2 diabetes and its management on the prognosis of patients with severe COVID -19

Impaired fasting glucose and diabetes are related to higher risks of complications and mortality among patients with coronavirus disease 2019

Newly diagnosed diabetes is associated with a higher risk of mortality than known diabetes in hospitalized patients with COVID -19

Managing type 2 diabetes mellitus during COVID-19 pandemic: the bittersweet

Baseline characteristics and risk factors for short-term outcomes in 132 COVID-19 patients with diabetes in Wuhan China: a retrospective study

Clinical analysis of risk factors for severe COVID-19 patients with type 2 diabetes

Hyperglycemia is a strong predictor of poor prognosis in COVID-19

Association of blood glucose control and outcomes in patients with COVID-19 and pre-existing type 2 diabetes

Clinical characteristics and outcomes of patients with diabetes and COVID-19 in association with glucoselowering medication

Association of diabetes mellitus with disease severity and prognosis in COVID-19: a retrospective cohort study

Influence of diabetes mellitus on the severity and fatality of SARS-CoV-2 (COVID-19) infection

Impaired glucose metabolism in patients with diabetes, prediabetes, and obesity is associated with severe COVID-19

Glycemic control before admission is an important determinant of prognosis in patients with coronavirus disease 2019

The clinical characteristics and outcomes of patients with diabetes and secondary hyperglycaemia with coronavirus disease 2019: a single-centre, retrospective, observational study in Wuhan

Type 1 diabetes in people hospitalized for COVID-19: new insights from the CORONADO study. Dia Care

Impact of glycemic control in diabetes mellitus on management of COVID-19 infection

Risks of and risk factors for COVID-19 disease in people with diabetes: a cohort study of the total population of Scotland

Risk factors of critical & mortal COVID-19 cases: a systematic literature review and meta-analysis

Impact of comorbidities and glycemia at admission and dipeptidyl peptidase 4 inhibitors in patients with type 2 diabetes with COVID-19: a case series from an academic hospital in Lombardy

Cardiometabolic multimorbidity is associated with a worse Covid-19 prognosis than individual cardiometabolic risk factors: a multicentre retrospective study (CoViDiab II)

Physiological and metabolic consequences of obesity

Inflammatory markers and control of type 2 diabetes mellitus

Comorbid diabetes results in immune dysregulation and enhanced disease severity following MERS-CoV infection

Phagocytic activity is impaired in type 2 diabetes mellitus and increases after metabolic improvement

Prevalence of residual inflammatory risk and associated clinical variables in patients with type 2 diabetes

CD31+ extracellular vesicles from patients with type 2 diabetes shuttle a miRNA signature associated with cardiovascular complications

Inflamm-aging: why older men are the most susceptible to SARS-CoV-2 complicated outcomes

Inflammatory cytokine concentrations are acutely increased by hyperglycemia in humans

Cardiovascular effects of gliptins

Effects of initiating insulin and metformin on glycemic control and inflammatory biomarkers among patients with type 2 diabetes

Elevated HbA1c levels in pre-Covid-19 infection increases the risk of mortality: A sistematic review and meta-analysis