Spotlight on hospital care for cirrhosis

Recent studies looked at improving hospital care for patients with cirrhosis.

The first study, published by the American Journal of Gastroenterology on Aug. 12, evaluated whether a cirrhosis admission order set would improve adherence to quality metrics and outcomes. Researchers at a hospital in Indiana developed and piloted an alert system in the electronic health record (EHR) that directed clinicians to an order set for patients who had or were likely to have cirrhosis. The study compared process measures for ascites, GI bleeding, and hepatic encephalopathy, as well as outcomes when the clinical decision support tool was used versus baseline performance before it was implemented. Two hundred two patients were included in the preimplementation cohort, 132 were included in the pilot cohort, and 133 were included in the implementation cohort.

The predictive alert correctly captured 81.8% of patients with cirrhosis admitted during the implementation period. Overall, implementation of the order set was associated with a significant increase in adherence to such process measures as diagnostic paracentesis (29.6% to 51.1%), low-sodium diet (34.3% to 77.8%), and involvement of social work (36.6% to 77.8%) (P<0.001 for all comparisons). In addition, length of intensive care and hospital stay (P<0.001) and in-hospital infection rates (P=0.002) decreased after the order set was implemented. Thirty- and 90-day readmission rates did not differ before and after implementation (P=0.897 and P=0.640, respectively). The researchers concluded that EHR-based clinical decision support tools can improve adherence to quality metrics for patients with cirrhosis. They called for additional studies with larger sample sizes to determine whether additional outcome measures could be improved.

The second study, published by the American Journal of Gastroenterology on Aug. 12, used data from the ATTIRE trial to examine whether antibiotics helped prevent hospital-acquired infection (HAI) in patients with cirrhosis. Patients in England, Wales, and Scotland who were hospitalized with cirrhosis but did not have infection at baseline were grouped into those who received an antibiotic prescription during hospitalization (other than rifaximin) and those who didn't. The researchers evaluated HAIs during the treatment period and mortality rates by using propensity score matching to account for differences in disease severity.

No differences in subsequent HAIs from days 3 to 15 were seen between the 203 patients who were prescribed antibiotics at study enrollment and the 360 who were not (19.2% vs. 20.3%; P=0.83). Antibiotic-treated patients had higher 28-day mortality rates (P=0.004), probably due to more severe disease. Among patients without infection at study enrollment, no differences in HAIs or mortality rates were seen between those who were taking rifaximin before admission (n=67) and those who were not (n=497). Six-month mortality rates did not differ between patients who received long-term antibiotic prophylaxis at discharge and those who did not.

The authors concluded that half of the antibiotics prescribed at study entry were given to patients who did not have a diagnosed infection and that the drugs did not decrease overall HAI risk or improve mortality rates. “These data support a policy of prompt de-escalation or discontinuation of empirical antibiotics guided by culture sensitivities at 24-48 hours after commencement if no infection and an improving patient,” they wrote.

Finally, a study published Aug. 15 by the Journal of Clinical Gastroenterology used the 2014 Nationwide Readmission Database to evaluate hospital outcomes of patients with cirrhosis and nonvariceal upper GI bleeding. ICD-9-CM codes were used to determine diagnoses and interventions. The primary outcomes were 30-day readmission rates, index admission mortality rates, health care utilization, and predictors of readmission and mortality using multivariable regression analysis. The study included 176,978 patients who were admitted with nonvariceal upper GI bleeding and were further stratified as having compensated cirrhosis (n=8,685) or decompensated cirrhosis (n=5,016), with patients who did not have cirrhosis included as controls (n=163,276).

The 30-day readmission rate was 20.8% for all patients with cirrhosis. Patients with compensated cirrhosis were more likely than those with decompensated cirrhosis or no cirrhosis to have an esophagogastroduodenoscopy (EGD) within one calendar day of admission (74.1%, 67.9%, and 69.4%, respectively). Hospitalizations were longer and costs of care were higher in patients with decompensated cirrhosis, and this group also had a higher index admission mortality rate (6.2%) than those with compensated cirrhosis (1.7%; P<0.001) or no cirrhosis (1.4%; P<0.001). An EGD performed more than one calendar day from admission was a predictor of 30-day readmission (odds ratio [OR], 1.21 [95% CI, 1.00 to 1.46]), while decompensated cirrhosis was a predictor of 30-day readmission and index admission mortality (ORs, 1.78 [95% CI, 1.54 to 2.06] and 3.68 [95% CI, 2.67 to 5.05], respectively).

The authors concluded that nonvariceal upper GI bleeding among patients with cirrhosis is associated with higher readmission rates, mortality rates, and health care utilization. “Early EGD is a modifiable predictor associated with reduced 30-day readmission rates. Although the other predictors of readmissions and mortality were nonmodifiable (cirrhosis compensation, age, number of other comorbidities, and hospital type), these variables can be helpful in identifying patients at a higher risk of poor outcomes at the time of their presentation to the hospital,” the authors concluded. “Early identification of these high-risk patients can be helpful in improving both clinical and health care system outcomes at a national level.”