Abstract
Background/Objectives: Histologic remission has emerged as a key treatment target in inflammatory bowel disease (IBD), but routine assessment requires repeated endoscopy and biopsies. Blood-based indices reflecting inflammation, nutritional status and atherogenic risk are inexpensive and widely available, yet their integrated contribution to histologic activity remains unclear. This study addresses this gap by simultaneously analyzing a broad panel of 44 variables-including nutritional status indicators, CBC-derived inflammation indices, and atherogenic lipid indices-in IBD patients. Methods: In this retrospective study, 100 patients with IBD (50 Crohn's disease [CD], 50 ulcerative colitis [UC]) without additional comorbidities and with concomitant histologic assessment were analyzed. Histologic activity was coded as active vs. remission. At the time of biopsy, the complete blood count, biochemistry and lipid profile were used to calculate immuno-nutritional indices (CONUT score, prognostic nutritional index), inflammatory indices (neutrophil-to-platelet ratio, platelet-to-lymphocyte ratio, lymphocyte-to-monocyte ratio [LMR], systemic immune-inflammation index, systemic immune-inflammation index, systemic inflammation response index [SIRI], aggregate index of systemic inflammation, C-reactive protein to albumin ratio) and atherogenic indices (atherogenic index of plasma [AIP], triglyceride-to-HDL cholesterol ratio). Variable selection was performed separately for CD and UC using least absolute shrinkage and selection operator (LASSO) regression and sparse partial least squares discriminant analysis (sPLS-DA). Independently associated predictors were then entered into multivariable logistic regression models, and their discriminative performance was evaluated using ROC analysis with bootstrap-derived 95% confidence intervals. Results: LASSO analysis identified a broadly similar systemic profile associated with histologic activity in CD and UC, dominated by the CONUT score, SIRI, AIP, LMR and red blood cell parameters, whereas demographic features and most routine biochemical markers were shrunk towards zero. Cross-validated AUCs for the LASSO models were 0.93 in CD and 0.87 in UC. sPLS-DA confirmed this pattern: CONUT, SIRI and AIP consistently showed the highest variable importance in projection scores and loadings on the first latent component. In multivariable regression, the CONUT score, SIRI and AIP remained independent predictors of histologic activity in CD, while hematocrit, CONUT score, SIRI and AIP were independently associated with histologic activity in UC. In ROC analysis, AUCs for CONUT, SIRI and AIP were 0.81, 0.89 and 0.87 in UC, and 0.72, 0.82 and 0.83 in CD, respectively. Conclusions: Histologic activity in IBD is characterized by a coupled systemic profile in which immuno-nutritional compromise (captured by CONUT) forms the core signal, supplemented by systemic inflammation (SIRI) and atherogenic dyslipidemia (AIP). These readily available blood-based indices may help to approximate histologic disease activity in clinical practice. However, considering that comorbid diseases may affect these indices, the strict exclusion criteria applied in this study may limit the generalizability of the findings among patients with IBD. Consequently, further validation in larger prospective cohorts is warranted.
-
Kapsamı
Uluslararası
-
Type
Hakemli
-
Index info
WOS.SCI
-
Language
English
-
Article Type
None