Potential conflict of interest: Nothing to report.
Predictors of outcomes in patients with ascites, hyponatremia, and renal failure†
Article first published online: 21 JUN 2013
Copyright © 2013 the American Association for the Study of Liver Diseases
Clinical Liver Disease
Special Issue: Cirrhosis and Kidney Function
Volume 2, Issue 3, pages 132–135, June 2013
How to Cite
Asrani, S. K. and Kamath, P. S. (2013), Predictors of outcomes in patients with ascites, hyponatremia, and renal failure. Clinical Liver Disease, 2: 132–135. doi: 10.1002/cld.203
- Issue published online: 21 JUN 2013
- Article first published online: 21 JUN 2013
- Manuscript Accepted: 22 MAR 2013
- Manuscript Revised: 25 FEB 2013
- Manuscript Received: 4 JAN 2013
Watch a video presentation of this article
Watch the interview with the author
The development of decompensated disease and complications of portal hypertension portends a poor prognosis for patients with cirrhosis. In comparison with the general population, patients with cirrhosis and compensated disease have a 5-fold increase in mortality, and patients with cirrhosis and decompensated disease have a 10-fold increase in mortality.1 Survival is lower at 1 (75% versus 87%) and 5 years (45% versus 67%) for persons with decompensated cirrhosis versus persons with compensated cirrhosis.1 The development of hyponatremia, ascites, and renal failure represents an unfortunate marker of disease severity in the natural history of cirrhosis.
The development of ascites portends poor survival. Among alcoholics studied in the Danish National Health registry (1993-2005), the 1-year survival rates were 80% (variceal bleeding), 71% (ascites), 51% (ascites and variceal bleeding), and 36% (hepatic encephalopathy alone or in combination).2 The presence of an infection, that is, spontaneous bacterial peritonitis (SBP), indicates a poorer outcome. In a recent meta-analysis,3 the overall median mortality rate for infected patients was 38%. In a subset of studies of SBP, the median mortality rate was 44% (1-year survival = 34%). Therefore, the prevention of infection remains a mainstay for decreasing mortality in patients with ascites. This includes primary prophylaxis for patients with cirrhosis who present with gastrointestinal bleeding and secondary prophylaxis for patients with a prior episode of SBP. In patients with a new diagnosis of ascites, a diagnostic paracentesis is imperative and serves as an important tool for predicting outcomes. Patients with low-protein ascites (<1.5 mg/dL) and specifically those with either severe liver disease (Child-Pugh score > 9 points and serum bilirubin level > 3 mg/dL) or renal dysfunction (serum creatinine level > 1.2 mg/dL, blood urea nitrogen level > 25 mg/dL, or serum sodium level < 130 mEq/L) benefit from primary antibiotic prophylaxis to prevent SBP (7% versus 61%), hepatorenal syndrome (HRS; 28% versus 41%), and 1-year mortality (40% versus 52%)4 (Fig. 1).
Hyponatremia is an independent predictor of 3-month and 1-year mortality for patients with liver disease.5 Hyponatremia is likely a marker for advanced renal dysfunction associated with cirrhosis that is not accurately captured by creatinine-based measures.6 Even after adjustments for Model for End-Stage Liver Disease (MELD) scores, a 1 mmol/L decrease in the serum sodium concentration between 125 and 140 mmol/L is associated with a 5% to 10% increase in mortality.5
Renal failure is common among patients with cirrhosis and is often due to several causes.7 In comparison with patients without renal failure, patients with cirrhosis and renal dysfunction have a significantly increased risk of death (odds ratio = 7.6, confidence interval = 5.4-10.8).8 The median survival rate for patients with cirrhosis and renal failure is 42% at 1 month and 37% at 1 year.8 Although renal failure itself is associated with a poor prognosis, recent data suggest that the cause of renal failure may help to predict the outcome. In a single-center study,9 the survival rate at 3 months was 73% for patients with cirrhosis and intrinsic renal disease, 46% for patients with volume-related renal failure, 31% for patients with infection-related renal failure, and 15% for patients with HRS (Fig. 2). The cause of renal failure was an independent predictor of mortality after adjustments for the MELD score and serum sodium. In a separate single-center study,10 the cause of renal failure was an important predictor of renal dysfunction and mortality 1 year after liver transplantation (Fig. 3). Patients with acute tubular necrosis (ATN) had the highest cumulative incidence of chronic kidney disease; the incidence of chronic kidney disease at 5 years was higher for patients with ATN versus patients with HRS (56% versus 16%). The presence of ATN at the time of liver transplantation was associated with a 7-fold increase in mortality 1 year after transplantation (odds ratio = 6.7, confidence interval = 1.97-22.8).10 The presence of ATN in unselected patients undergoing renal biopsy at the time of liver transplantation is a predictor of worse renal function 1 year after liver transplantation.11
In addition, the method of renal function measurement is important. The inclusion of serum creatinine is suboptimal for patients with cirrhosis.6, 12 The measured glomerular filtration rate (e.g., by iothalamate clearance measurement) is better at assessing prognosis.12 The role of cystatin C in the assessment of renal dysfunction in patients with cirrhosis is currently being evaluated. Furthermore, the role of urinary biomarkers as predictors of early renal dysfunction continues to evolve.
Prognostic Models (Table 1)
The Child-Pugh score continues to be an important tool for predicting the prognosis of patients with decompensated liver disease and ascites. The median 1-year survival rate is 95% for Child A cirrhosis, 80% for Child B cirrhosis, and 45% for Child C cirrhosis.15 Besides its subjective nature, the lack of inclusion of variables that assess renal function is a limitation of the Child-Pugh score.
|1 Point||2 Points||3 Points|
|Total bilirubin (mg/dL)||≤2||2–3||>3|
|Serum albumin (g/dL)||>3.5||2.8–3.5||<2.8|
|Ascites||None||Mild||Moderate to severe|
|Hepatic encephalopathy||None||Grade I-II||Grade III-IV|
|MELD score = 3.8 × loge Bilirubin + 11.2 × loge INR + 9.6 × loge Creatinine + 6.4|
|MELD-Na score = MELD score − Na − [0.025 × MELD × (140 − Na)] + 140|
|Refit MELD Score§14|
|Refit MELD score = 4.082 × loge Bilirubin + 8.485 × loge Creatinine + 10.671 × loge INR + 7.432|
|Refit MELD-Na Score∥14|
|Refit MELD-Na score = 4.258 × loge Bilirubin + 6.792 × loge Creatinine + 8.290 × loge INR + 0.652 × (140 − Na) − 0.194 × (140 − Na) × Bilirubin# + 6.327|
The MELD score is a commonly used metric of underlying disease severity.16 The benefits of the MELD score lie in its simplicity, use of a continuous scale, and incorporation of easily available and objective laboratory parameters. The MELD score has a high concordance for predicting the risk of death within 3 months (c statistic = 0.83) and 1 year (c statistic = 0.75). The individual complications of portal hypertension (e.g., ascites) do not significantly enhance the prognostic information.16 The MELD score captures the importance of renal failure as a marker of advanced liver disease. However, an obvious limitation is its inclusion of serum creatinine as reflective of renal dysfunction. A multivariate model that incorporates the calculated glomerular filtration rate and/or serum sodium is superior to the MELD score.6
As mentioned previously, serum sodium is likely a marker for advanced renal dysfunction that is not accurately captured by creatinine-based measures and adds prognostic value [Model for End-Stage Liver Disease with sodium (MELD-Na)]. The incorporation of the MELD-Na score (c statistic = 0.88) as a predictor of mortality for persons with advanced liver disease implies that 88 out of 100 times, the MELD-Na score can accurately predict that the patient with a higher score in a pair of patients with cirrhosis has the higher risk of death within 3 months.
The development of hyponatremia, ascites, and renal failure represents an unfortunate milestone in the natural history of cirrhosis. First, patients with low-protein ascites and severe liver disease or renal dysfunction have a poor prognosis and may benefit from primary antibiotic prophylaxis. Second, serum sodium is a marker for advanced renal dysfunction that is not accurately captured by serum creatinine–based measures and adds prognostic value. Third, the cause of renal failure independently predicts mortality after adjustments for the MELD score and serum sodium. Direct measurements of renal function are better than creatinine-based estimates of renal function and prognosis. The role of urinary markers continues to evolve. Fourth, prognostic scores used in the assessment of patients with cirrhosis include the Child-Pugh score, MELD score, MELD-Na score, and newer iterations that refit the coefficients of the MELD score.
- 11Acute tubular necrosis is present in the majority of patients undergoing liver transplantation [abstract]. Hepatology 2011; 54: 159A., , , , , , et al.
- 12Inaccuracies of creatinine and creatinine-based equations in candidates for liver transplantation with low creatinine: impact on the Model for End-Stage Liver Disease score. Liver Transpl 2010; 16: 1169–1177., , , , , , et al.