Ethnicity influences phenotype and clinical outcomes: Comparing a South American with a North American inflammatory bowel disease cohort

Abstract

Inflammatory bowel disease (IBD), including ulcerative colitis (UC) and Crohn disease (CD), has emerged as a global disease with an increasing incidence in developing and newly industrialized regions such as South America. This global rise offers the opportunity to explore the differences and similarities in disease presentation and outcomes across different genetic backgrounds and geographic locations. Our study includes 265 IBD patients. We performed an exploratory analysis of the databases of Chilean and North American IBD patients to compare the clinical phenotypes between the cohorts. We employed an unsupervised machine-learning approach using principal component analysis, uniform manifold approximation, and projection, among others, for each disease. Finally, we predicted the cohort (North American vs Chilean) using a random forest. Several unsupervised machine learning methods have separated the 2 main groups, supporting the differences between North American and Chilean patients with each disease. The variables that explained the loadings of the clinical metadata on the principal components were related to the therapies and disease extension/location at diagnosis. Our random forest models were trained for cohort classification based on clinical characteristics, obtaining high accuracy (0.86 = UC; 0.79 = CD). Similarly, variables related to therapy and disease extension/location had a high Gini index. Similarly, univariate analysis showed a later CD age at diagnosis in Chilean IBD patients (37 vs 24; P = .005). Our study suggests a clinical difference between North American and Chilean IBD patients: later CD age at diagnosis with a predominantly less aggressive phenotype (39% vs 54% B1) and more limited disease, despite fewer biological therapies being used in Chile for both diseases.

Figure 1.

Analysis of clinical metadata of UC patients. To explore the relation between centers and clinical metadata (disease phenotype, treatments, disease activity, location disease, and complications), we used a PCA dimension reduction technique and then plotted the results in a 2D plot. These results show that 2 groups arise from the clinical data, one for Chilean patients and another for American patients. Also, PCA component variance explanation is shown. (A) PCA in UC. The patients were color-coded by cohort (Chilean blue dots and North American red dots). The first 5 PCA components represent the 50% cumulative variance. (B) Variable importance PC1. (C) Variable importance PC2. 2D = 2 dimensional, PC1 = first principal component, PCA = principal component analysis, UC = ulcerative colitis.

Figure 2.

Analysis of clinical metadata of CD patients. Analysis of clinical metadata of CD patients. (A) PCA in CD. The patients are color-coded according to cohort: Chilean blue dots and North American red dots. The first 5 PCA components represent the 50% cumulative variance. (B) Variable importance PC1. (C) Variable importance PC2. * current_anti_12_23_CD = current use of ant IL-12 and IL-23; Grand_total = Total score SES-CD. CD = Crohn disease, IL = interleukin, PC1 = first principal component, PC2 = second principal component, PCA = principal component analysis, SES-CD = simple endoscopic score for Crohn’s disease, UC = ulcerative colitis.

Figure 3.

Dimensional reduction of IBD patient clinical data. (A–D) Ulcerative colitis patients’ data represented using t-SNE, UMAP, MDS, Kernel PCA (left to right). (E–H) Crohn Disease patients represented by the same algorithms. Patient Cohort or “center” is represented in blue for Chile and yellow for UCSD. nUC = 173, nCD = 92. IBD = inflammatory bowel disease, MDS = multidimensional scaling, PCA = principal component analysis, t-SNE = t-distributed stochastic neighbor embedding, UMAP = uniform manifold approximation, and projection, UCSD = University California San Diego.

Figure 4.

Variable importance for cohort prediction. To assess the importance of each variable for the random forest model, we plotted a horizontal bar graph, where the relative importance of each variable measured by the Gini index is shown. (A) the Gini index shows the main variables to predict the cohort (Chilean vs North American) in UC patients were related to use of biological therapies, aminosalicylates, inflammatory parameters, duration of disease, and UC extension. On the contrary, (B) shows that in CD patients the main variables were related to the use of therapies, duration of disease, inflammatory parameters, clinical activity score, and localization of the disease. *current_anti_12_23_CD = current use of ant IL-12 and IL-23; Grand_total = Total score SES-CD. CD = Crohn disease, IL = interleukin, SES-CD = simple endoscopic score for Crohn’s disease, UC = ulcerative colitis.