Longitudinal Cluster Analysis of Hemodialysis Patients with COVID-19 in the Pre-Vaccination Era

Area of Science:

• Nephrology
• Infectious Diseases
• Clinical Data Science

Background:

• Coronavirus disease 2019 (COVID-19) presents heterogeneously in hemodialysis (HD) patients.
• Effective patient stratification is needed to predict COVID-19 severity and outcomes in this vulnerable population.

Purpose of the Study:

• To develop and validate a method for stratifying hemodialysis patients based on COVID-19 clinical and laboratory data.
• To identify distinct patient clusters associated with different disease trajectories and complication risks.

Main Methods:

• Collected clinical and laboratory data from two cohorts of hemodialysis patients diagnosed with COVID-19.
• Utilized a linear mixed-effects model (LME) with baseline and longitudinal data to define patient clusters.
• Validated the LME model and cluster assignments in an independent cohort.

Main Results:

• Two distinct clusters (cl1 and cl2) were identified, characterized by differences in inflammatory markers and immune cell populations.
• Cluster 1 (cl1) patients exhibited a milder COVID-19 presentation, with lower disease activity, hospitalization rates, mortality, and oxygen requirements.
• The clustering analysis successfully stratified patients, enabling the identification of those at higher risk for complications.

Conclusions:

• Longitudinal data clustering provides a viable strategy for stratifying COVID-19 patients on hemodialysis.
• This approach aids in identifying patients at high risk for severe disease and complications.
• The developed strategy shows potential applicability across various clinical settings for managing COVID-19 in HD patients.