Data-driven identification of temporal glucose patterns in a large cohort of nondiabetic patients with COVID-19 using time-series clustering

Abstract

Objective Hyperglycemia has emerged as an important clinical manifestation of coronavirus disease 2019 (COVID-19) in diabetic and non-diabetic patients. Whether these glycemic changes are specific to a subgroup of patients and persist following COVID-19 resolution remains to be elucidated. This work aimed to characterize longitudinal random blood glucose in a large cohort of non-diabetic patients diagnosed with COVID-19. Materials and Methods De-identified electronic medical records of 7,502 patients diagnosed with COVID-19 without prior diagnosis of diabetes between January 1st and November 18th, 2020 were accessed through the TriNetX Research Network. Glucose measurements, diagnostic codes, medication codes, laboratory values, vital signs, and demographics were extracted before, during, and after COVID-19 diagnosis. Unsupervised time-series clustering algorithms were trained to identify distinct clusters of glucose trajectories. Cluster associations were tested for demographic variables, COVID-19 severity, glucose-altering medications, glucose values, and new-onset diabetes diagnoses. Results Time-series clustering identified a low-complexity model with 3 clusters and a high-complexity model with 19 clusters as the best-performing models. In both models, cluster membership differed significantly by death status, COVID-19 severity, and glucose levels. Clusters membership in the 19 cluster model also differed significantly by age, sex, and new-onset diabetes mellitus. Discussion and Conclusion This work identified distinct longitudinal blood glucose changes associated with subclinical glucose dysfunction in the low-complexity model and increased new-onset-diabetes incidence in the high complexity model. Together, these findings highlight the utility of data-driven techniques to elucidate longitudinal glycemic dysfunction in patients with COVID-19 and provide clinical evidence for further evaluation of the role of COVID-19 in diabetes pathogenesis. Lay Summary Hyperglycemia is defined as elevated blood glucose measurements and are common in diabetic patients. Recent findings suggest that patients diagnosed with COVID-19 may experience elevated blood glucose levels during COVID-19 infection. Whether blood glucose levels remain elevated after COVID-19 infection remains poorly understood. This study aimed to identify how patterns of blood glucose levels before, during, and after COVID-19 change. This work analyzed blood glucose levels from 2,059 patients diagnosed with COVID-19 and used machine learning to identify different patterns of blood glucose changes. We found patterns demonstrating an overall increase, overall decrease, temporal increase, and temporary decrease in blood glucose levels. Some of these patterns were associated with COVID-19 severity and proportion of patients with new-onset diabetes. These findings demonstrate the usefulness of machine learning in understanding glucose changes following COVID-19 diagnosis and indicate that more research is needed to understand if blood glucose monitoring in COVID-19 patients should be routinely performed.

Prevalence and risk of new-onset diabetes mellitus after COVID-19: a systematic review and meta-analysis

“…In three studies, a matched-cohort study design was used to compare diabetes incidence in a group of participants with previous SARS-CoV-2 infection with a matched control group with no previous infection. In such cases people with prevalent diabetes were specifically excluded at study entry ( 4 , 30 , 34 ). These studies showed that in the cohort of people with previous SARS-CoV-2 infection the incidence of diabetes was higher than the corresponding historical or contemporary non-COVID-19 control cohort, suggesting that infected people might be at higher risk of developing diabetes than people who was not infected.…”

“…Two articles were further excluded because no quantitative data were available after contacting the Authors (16,17). Finally, 20 original paper were included in the review (4,(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35)(36), with the main characteristics of the studies displayed in Table 1. The detailed selection process is depicted in Figure 1.…”

“…diabetes were specifically excluded at study entry (4,30,34). These studies showed that in the cohort of people with previous SARS-CoV-2 infection the incidence of diabetes was higher than the corresponding historical or contemporary non-COVID-19 control cohort, suggesting that infected people might be at higher risk of developing diabetes than people who was not infected.…”