Gut Microbiome-Based Diagnostic Model to Predict Coronary Artery Disease

Abstract

In the present study, we aimed to characterize gut microbiome and develop a gut microbiome-based diagnostic model in patients with coronary artery disease (CAD). Prospectively, we collected 309 fecal samples from Central China and Northwest China and carried out the sequencing of the V3−V4 regions of the 16S rRNA gene. The gut microbiome was characterized, and microbial biomarkers were identified in 152 CAD patients and 105 healthy controls (Xinjiang cohort, n = 257). Using the biomarkers, we constructed a diagnostic model and validated it externally in 34 CAD patients and 18 healthy controls (Zhengzhou cohort, n = 52). Fecal microbial diversity was increased in CAD patients compared to that in healthy controls (P = 0.021). Phylum Bacteroidetes was increased in CAD patients versus healthy controls (P = 0.001). Correspondingly, 48 microbial markers were identified through a 10-fold cross-validation on a random forest model, and an area under the curve (AUC) of 87.7% (95% CI: 0.832 to 0.916, P < 0.001) was achieved in the Xinjiang cohort (development cohort, n = 257). Notably, an AUC of 90.4% (95% CI: 0.848 to 0.928, P < 0.001) was achieved using combined analysis of gut microbial markers and clinical variables. This model provided a robust tool for the prediction of CAD. It could be widely employed to complement the clinical assessment and prevention of CAD.

The Types and Proportions of Commensal Microbiota Have a Predictive Value in Coronary Heart Disease

“…Four studies were excluded due to insufficient information pertaining to the inclusion criteria, and a second article by the same authors was found to be a repetitive report based on a partial dataset. In total, 16 publications, reporting on 16 cohort studies, were selected for inclusion in the analysis [ 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 ].…”

Gut Microbiome-Based Diagnostic Model to Predict Diabetes Mellitus

“…In addition, for the screening of marker bacteria, we employed a mechanical learning model to screen for important bacteria. For the calculation of the probability of disease (POD) index, we performed logarithmic transformation on the bacterial abundance, and combined logistics regression to calculate the probability as the POD index [20].…”

“…Subsequently, we performed logarithmic transformation on the microbial abundance of these 12 genera microorganisms. We calculated the POD value in combination with logistics regression [20], and established the nomogram diagnostic model in combination with age, gender (if female), and clinical variable indicators such as the denial from the diagnosis of hypertension and fasting blood glucose (See Fig. 7A).…”

Pet-Human Gut Microbiome Host Classifier Using Data from Different Studies

“…The gut microbiota, despite its complexity and great variation between individuals, was shown to be predictive of various intestinal diseases and conditions, such as irritable bowel syndrome (IBS) [ 3 ], Crohn’s disease [ 4 , 5 ], and colorectal cancer [ 6 ]. Interestingly, the composition of the gut microbiome also predicts some non-intestinal illnesses, such as coronary artery disease [ 7 ], liver fibrosis [ 8 ], metabolic diseases/obesity [ 9 ], insomnia [ 10 ], and bipolar depression [ 11 ]. Another classification setting is the detection of contamination in samples.…”

“…Random forest (RF) algorithm [24] was chosen for the classification task. RFs are highly used machine learning algorithms for microbiome classification [3][4][5][6][7][8][9][10][11] due to the limited number of model parameters and simple results interpretation.…”