Predictors of outcomes in patients with ascites, hyponatremia, and renal failure

Authors

  • Sumeet K. Asrani M.D.,

    1. Division of Gastroenterology and Hepatology, Mayo Clinic College of Medicine, Rochester, MN
    2. Baylor University Medical Center, Dallas, TX
    Search for more papers by this author
  • Patrick S. Kamath M.D.

    Corresponding author
    1. Division of Gastroenterology and Hepatology, Mayo Clinic College of Medicine, Rochester, MN
    • Division of Gastroenterology and Hepatology, Mayo Clinic College of Medicine, 200 First Street Southwest, Rochester, MI 55905
    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

Abstract

Watch a video presentation of this article

Watch the interview with the author

Answer questions and earn CME

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.

Ascites

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).

Figure 1.

Primary prophylaxis for the prevention of SBP among patients with cirrhosis or significant renal dysfunction. Reprinted with permission from Gastroenterology.4 Copyright 2007, Elsevier.

Abbreviations
AKI

acute kidney injury

ATN

acute tubular necrosis

HRS

hepatorenal syndrome

INR

international normalized ratio

MELD

Model for End-Stage Liver Disease

MELD-Na

Model for End-Stage Liver Disease with sodium

Na

serum sodium concentration

SBP

spontaneous bacterial peritonitis

Hyponatremia

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

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

Figure 2.

Etiology of renal failure as a predictor of mortality among patients with cirrhosis. Reprinted with permission from Gastroenterology.9 Copyright 2011, Elsevier.

Figure 3.

Etiology of pre–liver transplant renal failure as a predictor of mortality after liver transplantation. Reprinted with permission from Liver Transplantation.10 Copyright 2012, American Association for the Study of Liver Diseases.

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.

Table 1. Prognostic Scores for the Assessment of Patients With Cirrhosis
 Child-Pugh Score*
 1 Point2 Points3 Points
  • Abbreviations: INR, international normalized ratio; Na, serum sodium concentration.

  • *

    The Child-Pugh stage is determined as follows: (A) 5 to 6 points, (B) 7 to 9 points, and (C) 10 to 15 points.

  • The bilirubin and creatinine levels are measured in milligrams per deciliter.

  • The lower and upper boundaries for Na are 125 and 140 mmol/L, respectively.

  • §

    The lower boundary for bilirubin is 1 mg/dL. The lower and upper boundaries for creatinine are 0.8 and 3 mg/dL, respectively. The lower and upper boundaries for INR are 1 and 3, respectively. Renal replacement therapy is equal to 3 mg/dL.

  • The lower boundary for bilirubin is 1 mg/dL. The lower and upper boundaries for creatinine are 0.8 and 3 mg/dL, respectively. The lower and upper boundaries for INR are 1 and 3, respectively. Renal replacement therapy is equal to 3 mg/dL. The lower and upper boundaries for bilirubin# are 1 and 20 mg/dL, respectively.

Total bilirubin (mg/dL)≤22–3>3
Serum albumin (g/dL)>3.52.8–3.5<2.8
INR<1.71.7–2.3> 2.3
AscitesNoneMildModerate to severe
Hepatic encephalopathyNoneGrade I-IIGrade III-IV
MELD Score13
MELD score = 3.8 × loge Bilirubin + 11.2 × loge INR + 9.6 × loge Creatinine + 6.4
MELD-Na Score5
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 Score14
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.

Recently, attempts have been made to reweigh the components of the MELD score with a contemporary data set, and they have improved the predictive ability of MELD-based scores consistently to a c statistic > 0.87.14, 17 Further validation is awaited.

Summary

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.

Ancillary