Lower skeletal muscle attenuation and high visceral fat index are associated with complicated disease in patients with Crohn's disease: An exploratory study
2017
Cravo, Marília L. | Velho, Sónia | Torres, Joana | Costa Santos, Maria Pia | Palmela, Carolina | Cruz, Rita | Strecht, João | Maio, Rui | Baracos, Vickie
The prognostic value of body composition analysis in patients with Crohn's disease (CD) is poorly explored. The aims of the present study were to assess fat and skeletal muscle compartments including muscle radiation attenuation (MA) in patients with CD, and to analyze its predictive value to identify complicated phenotypes.Seventy one patients with CD who have had an abdominal CT within one month of clinical, laboratory, and endoscopic evaluation were included. Skeletal muscle area (SMA) and index (SMI), visceral fat area (VFA) and index (VFI), subcutaneous fat area (SFA), and mean MA were measured using appropriate software. Sarcopenia, as defined by Martin's criteria was assessed. Montreal classification was used to characterize disease phenotype.Mean MA was lower in patients >40 years (p = 0.001), L2 (p = 0.09) and stricturing/penetrating disease (p = 0.03) whereas SMA and SMI were significantly lower in patients with positive C-reactive protein and previous hospital admissions (p < 0.01). On multivariate analysis, higher MA was protective against the complicated disease phenotype (stricturing/penetrating disease and/or previous surgeries) (OR 0.81; p = 0.002) whereas a high visceral fat index increased such risk (OR 26.1; p = 0.02). A ROC curve showed a 82.4% sensibility, 90.3% specificity, 17.6% positive predictive value, 9.7% negative predictive value and an area under the curve (AUC) of 0.91 for body composition analysis to predict complicated disease.A lower muscle attenuation and a high visceral fat index seem to be associated with more severe phenotypes in patients with CD.
Показать больше [+] Меньше [-]Ключевые слова АГРОВОК
Библиографическая информация
Эту запись предоставил National Agricultural Library