Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2024 May 13:12:e17340.
doi: 10.7717/peerj.17340. eCollection 2024.

A retrospective prognostic evaluation using unsupervised learning in the treatment of COVID-19 patients with hypertension treated with ACEI/ARB drugs

Liye Ge  1 , Yongjun Meng  1 , Weina Ma  1 , Junyu Mu  2
Affiliations

A retrospective prognostic evaluation using unsupervised learning in the treatment of COVID-19 patients with hypertension treated with ACEI/ARB drugs

Liye Ge et al. PeerJ. .

Abstract

Introduction: This study aimed to evaluate the prognosis of patients with COVID-19 and hypertension who were treated with angiotensin-converting enzyme inhibitor (ACEI)/angiotensin receptor B (ARB) drugs and to identify key features affecting patient prognosis using an unsupervised learning method.

Methods: A large-scale clinical dataset, including patient information, medical history, and laboratory test results, was collected. Two hundred patients with COVID-19 and hypertension were included. After cluster analysis, patients were divided into good and poor prognosis groups. The unsupervised learning method was used to evaluate clinical characteristics and prognosis, and patients were divided into different prognosis groups. The improved wild dog optimization algorithm (IDOA) was used for feature selection and cluster analysis, followed by the IDOA-k-means algorithm. The impact of ACEI/ARB drugs on patient prognosis and key characteristics affecting patient prognosis were also analysed.

Results: Key features related to prognosis included baseline information and laboratory test results, while clinical symptoms and imaging results had low predictive power. The top six important features were age, hypertension grade, MuLBSTA, ACEI/ARB, NT-proBNP, and high-sensitivity troponin I. These features were consistent with the results of the unsupervised prediction model. A visualization system was developed based on these key features.

Conclusion: Using unsupervised learning and the improved k-means algorithm, this study accurately analysed the prognosis of patients with COVID-19 and hypertension. The use of ACEI/ARB drugs was found to be a protective factor for poor clinical prognosis. Unsupervised learning methods can be used to differentiate patient populations and assess treatment effects. This study identified important features affecting patient prognosis and developed a visualization system with clinical significance for prognosis assessment and treatment decision-making.

Keywords: ACEI; ARB; COVID-19; Hypertension; Prognostic evaluation; Unsupervised learning.

PubMed Disclaimer

Conflict of interest statement

The authors declare there are no competing interests.

Figures

Figure 1
Figure 1. The technical flowchart of this study.
(1) Data collection: First, we collected the required data from appropriate sources, which may include demographic information, clinical indicators, laboratory test results, and other content. (2) Feature extraction: Next, we process the collected data and extract features related to the research purpose. (3) Improved algorithm and synchronous optimization: To improve the initial point selection dependency of the traditional k-means algorithm, we introduce the swarm intelligence algorithm and conduct synchronous optimization to achieve the task of feature filtering and initial point optimization. (4) Feature selection and cluster analysis: After optimization, we further conducted feature selection to identify features that had a significant impact on the prognosis. Next, we use the improved k-means algorithm to cluster the data and divide the samples into different clusters. (5) Model training and evaluation: Based on clustering analysis, we use the selected features to train the prediction model and evaluate the model to evaluate its performance and predictive ability. The cases (80 cases) that meet the inclusion criteria after screening of COVID-19 patients (200 cases) are called residual samples. Residual samples refer to the leftover samples after the initial analysis or processing steps have been completed.
Figure 2
Figure 2. Performance analysis of the IDOA on 23 test functions.
This figure shows the results of 23 common test functions used in the performance evaluation of the improved IDOA. The evaluation results indicate that the IDOA outperforms the control method in terms of convergence speed and global optimal solution acquisition ability. The convergence speed refers to the speed at which an algorithm reaches the global optimal solution from its initial state, while the ability to obtain the global optimal solution indicates whether the algorithm can find the optimal solution to the problem. The figure shows that the IDOA algorithm converges faster and has better global optimal solution acquisition ability for most test functions.
Figure 3
Figure 3. Schematic diagram of synchronous optimization.
This process includes several key steps, including initialization, feature filtering, dimensionality determination of clustering center points, and merging and optimizing features and center points. From initialization to feature filtering, dimensionality determination of clustering center points, and then to the merging and optimization of features and center points, each step plays an important role, ultimately achieving optimization of clustering results.
Figure 4
Figure 4. Two-dimensional scatter plot after PCA dimensionality reduction.
Figure 5
Figure 5. Ranking of feature importance.
The 17 statistically significant characteristics in the feature difference analysis: age, whether to use ACEI/ARB, hypertension classification, COVID-19 classification, mulbsta score, lymphocytes, CRP, IL-6, lactate dehydrogenase, D-2 polymer, oxygenation, PCT, AST, creatinine, NT proBNP, hypersensitive sarcocalcin I, gene O were numbered Q1-Q17 respectively, and the importance of the features was sorted using the ReliefF algorithm.
Figure 6
Figure 6. Visualization system interface.
(A) Example of a patient with a good prognosis; (B) example of a patient with a poor prognosis.
See this image and copyright information in PMC

Similar articles

See all similar articles

References

    1. Angeli F, Reboldi G, Trapasso M, Zappa M, Spanevello A, Verdecchia P. COVID-19, vaccines and deficiency of ACE2 and other angiotensinases. Closing the loop on the Spike effect. European Journal of Internal Medicine. 2022;103:23–28. doi: 10.1016/j.ejim.2022.06.015. - DOI - PMC - PubMed
    1. Aslam M. Chi-square test under indeterminacy: an application using pulse count data. BMC Medical Research Methodology. 2021;21:201. doi: 10.1186/s12874-021-01400-z. - DOI - PMC - PubMed
    1. Asselah T, Durantel D, Pasmant E, Lau G, Schinazi RF. COVID-19: discovery, diagnostics and drug development. Journal of Hepatology. 2021;74:168–184. doi: 10.1016/j.jhep.2020.09.031. - DOI - PMC - PubMed
    1. Bauer A, Schreinlechner M, Sappler N, Dolejsi T, Tilg H, Aulinger BA, Weiss G, Bellmann-Weiler R, Adolf C, Wolf D, Pirklbauer M, Graziadei I, Gänzer H, Von Bary C, May AE, Wöll E, Von Scheidt W, Rassaf T, Duerschmied D, Brenner C, Kääb S, Metzler B, Joannidis M, Kain H-U, Kaiser N, Schwinger R, Witzenbichler B, Alber H, Straube F, Hartmann N, Achenbach S, Von Bergwelt-Baildon M, Von Stülpnagel L, Schoenherr S, Forer L, Embacher-Aichhorn S, Mansmann U, Rizas KD, Massberg S, Bantkowiak M, Baur G, Baylacher M, Beaucamp M, Berger M, Besch L, Brunner S, Budweiser S, Bugger H, Coletti R, Dorwarth U, Egresits J, Eiffener E, Faul C, Finkenstedt A, Gatos K, Gauchel N, Gindele F, Grander W, Gunschl M, Hartig F, Hecht M, Heer T, Heger L, Hentrich M, Horvath L, Keta D, Kiechl S, Kirchmaier R, Klein A, Klemm M, Kolesnik E, König A, Kossmann HC, Kropacek J, Lanser L, Lother A, Löw A, Mahabadi A-A, Malleier S, Mayer G, Müller C, Müller-Wieland D, Nagel B, Neuwirt H, Olivier C, Raunegger T, Reindl M, Reinstadler S, Riesinger L, Schäffner M, Schier J, Schock J, Schönherr P, Schulz M, Schütz T, Schwarz J, Siebermair J, Siry M, Spaur A, Sturm W, Tessadri K, Theurl F, Theurl M, Thommes L, Tiller C, Toifl M, Totzeck M, Von Zur Mühlen H, Vonderlin N, Wakili R, Wendtner C, Wenner F, Wimmert-Roidl D, Zabernigg A. Discontinuation versus continuation of renin-angiotensin-system inhibitors in COVID-19 (ACEI-COVID): a prospective, parallel group, randomised, controlled, open-label trial. The Lancet Respiratory Medicine. 2021;9:863–872. doi: 10.1016/S2213-2600(21)00214-9. - DOI - PMC - PubMed
    1. Bhandari S, Ives N, Brettell EA, Valente M, Cockwell P, Topham PS, Cleland JG, Khwaja A, El Nahas M. Multicentre randomized controlled trial of angiotensin-converting enzyme inhibitor/angiotensin receptor blocker withdrawal in advanced renal disease: the STOP-ACEi trial. Nephrology, Dialysis, Transplantation. 2016;31:255–261. doi: 10.1093/ndt/gfv346. - DOI - PMC - PubMed

Grants and funding

This work was supported by the Shanghai Key Specialty Project of Clinical Pharmacy (No.YXZDZK-01), the Nature Science Foundation of Jiading District, the Shanghai (No.JDKW-2021-0043) and the Shanghai University of Medicine and Health Sciences Clinical Research Centre for Metabolic Vascular Diseases Project (No.20MC2020004). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.