Background and aims
Skeletal muscle index (SMI) from computed tomography (CT) reliably assesses sarcopenia,
however, it is expensive and involves serial radiation exposure. Phase angle (PhA)
from bioimpedance analysis (BIA) is a noninvasive, low cost, bedside nutritional tool
used to monitor changes to nutritional interventions. We aimed to compare the performance
of PhA with SMI to assess sarcopenia in cirrhosis.
Ambispective cohort study. Consecutive patients with cirrhosis and available images
from abdominal CT scan were included. Monofrequency BIA was performed within two weeks
CT. Spearman´s correlation, ROC curve, and survival analysis with Kaplan-Meier, Cox
and competing-risk regression were performed.
136 patients were included with a mean age of 54.5 years (60% female). Most had decompensated
disease (66%) with ascites in 47%, and a mean MELD of 14 ± 6. We found positive correlations
between SMI and PhA (r=0.58, p<0.001), irrespective of the presence of ascites. The
AUROC of PhA-sarcopenia in all patients was 0.702; (0.748 in males,0.677 in females).
The best cutoffs of PhA for diagnosing sarcopenia were ≤5.6° in males and ≤5.4° in
females. SMI and PhA were significantly associated with survival in Kaplan-Meier curves.
In multivariable analyses, SMI was outperformed by age and MELD, whereas PhA remained
independently associated with mortality. Considering transplantation as a competing
risk, regression analysis showed both SMI and PhA to be independent predictors of
mortality (sHR:0.95 [0.90-0.99] and sHR:0.61 [0.42-0.88]).
PhA moderately correlates with SMI for the identification of sarcopenia in patients
with cirrhosis. However, its prognostic accuracy is comparable to that of SMI, and
it is not influenced by ascites.