Published online Jun 21, 2021. doi: 10.3748/wjg.v27.i23.3372
Peer-review started: January 28, 2021
First decision: April 5, 2021
Revised: April 14, 2021
Accepted: May 20, 2021
Article in press: May 20, 2021
Published online: June 21, 2021
Processing time: 140 Days and 18.4 Hours
The acute-on-chronic liver failure (ACLF) patients have a high short-term mortality rate, and the severity evaluation of ACLF is necessary for prognostication. The prevalence of diabetic mellitus (DM) has increased significantly in recent decades, resulting in coexistence of DM and chronic liver disease be common. There were very few studies that have evaluated the association between DM and ACLF. Therefore, the association between type 2 DM and the prognosis of patients with ACLF remained unclear.
Previous studies suggested that DM increased higher incidences of complications in chronic liver disease, included encephalopathy, spontaneous bacterial peritonitis, and hepatocellular carcinoma, but it was not clear whether DM would similarly affect ACLF patients. We needed further to evaluate the association between DM and the prognosis of ACLF patients in order to explore the feasibility of using DM as a prognostic indicator in ACLF patients.
We aimed to examine the effect of DM on the prognosis of patients with ACLF and established a predictive model to predict the risk of mortality for ACLF patients. A well-established prognostic predictive model was extremely important, which was helpful for early intervention of ACLF patients as well as improved survival rate.
Clinical data from 222 ACLF patients were retrospectively collected and analyzed between July 2013 and July 2020 from the Department of Hepatology Research Institute of the First Affiliated Hospital, Fujian Medical University. The patients were categorized into two groups depending on whether they had DM or not, and complications of ACLF were documented during treatment. Values of laboratory parameters, complication rates, and hospital mortality rates were compared between two groups. The Cox proportional hazard regression analysis was used to estimate hazard ratio (HR) for the association between DM and all-cause mortality. DM was independent risk factor to predict mortality in ACLF patients. Further, the nomogram and forest plot were built in terms of results of Cox regression analyses. A survival nomogram model was constructed to predict the probability of 1- and 2-mo survival for ACLF patients.
The prognosis of ACLF patients was significantly correlated with DM in univariate (HR = 2.4, P < 0.001) and multivariable analysis (HR = 3.17, P < 0.001). The ACLF patients with DM tended to have a high mortality rate (P < 0.001). Survival curves showed that the cumulative survival rate of ACLF patients was significantly distinguished between DM and non-DM (P = 0.00019). The survival time of ACLF patients with non-DM was longer than the patients with DM. As a good prognostic indicator, DM was helpful to predict the outcome of ACLF patients. The incidence of infection was higher in DM patients than non-DM counterparts (P = 0.038). Among all possible infections, spontaneous peritonitis and pulmonary infection should be considered a priority for ACLF patients. ACLF patients’ survival nomogram could predict the probability of 1- and 2-mo survival, which would be helpful to provide early clinical intervention. Because of the retrospective design of the study, we needed to assess further the impact of DM on the prognosis in ACLF patients by a prospective cohort study.
Our study found a significant association between DM and the prognosis of ACLF patients in China. The survival time of ACLF patients with non-DM was longer than that of patients with DM. The ACLF patients with DM had higher incidence of hospital mortality and infection than those without DM. DM was an independent risk factor affecting the prognosis of ACLF patients.
This was not a multicenter study and the number of DM patients with ACLF was relatively small. Therefore, a multi-center prospective cohort study would evaluate further the impact of DM on the prognosis in ACLF patients. Meanwhile, a noninvasive model should be established based on an extensive clinical database to predict effectively the survival time of ACLF patients.