ICD-10-CM Code for Ascites Due to Alcoholic Cirrhosis
The ICD-10-CM code for ascites due to alcoholic cirrhosis is K70.31 (Alcoholic cirrhosis of liver with ascites).
Code Structure and Application
The K70.31 code specifically captures the combination of alcoholic cirrhosis and ascites as a single diagnostic entity. This is the most precise code for this clinical scenario, as it directly links the etiology (alcohol) with the liver pathology (cirrhosis) and the complication (ascites) 1.
Alternative Coding Considerations
While K70.31 is the primary code, you should be aware of related codes that may appear in administrative data:
- K70.11: Alcoholic hepatitis with ascites (used when hepatitis is the predominant feature rather than established cirrhosis)
- R18.8: Other ascites (non-specific, should NOT be used when the etiology is known to be alcoholic cirrhosis)
- K71.51: Toxic liver disease with chronic active hepatitis with ascites (for drug-induced cases, not alcohol)
Important Coding Pitfalls
Do not assume all ascites in alcoholic patients is from alcoholic liver disease 2. The underlying cause must be confirmed through:
- Diagnostic paracentesis with ascitic fluid analysis
- Calculation of serum-ascites albumin gradient (SAAG ≥1.1 g/dL confirms portal hypertension) 3
- Exclusion of other causes (malignancy, cardiac failure, tuberculosis)
Avoid using R18.8 when the specific etiology is known. Research demonstrates that R18.8 alone has poor sensitivity for identifying cirrhotic ascites and should be reserved for cases where the underlying cause is truly uncertain 1, 4.
Documentation Requirements
For accurate coding, clinical documentation should include:
- Confirmed history of chronic alcohol use
- Evidence of cirrhosis (imaging, biopsy, or clinical decompensation)
- Presence of ascites (physical examination, imaging, or paracentesis)
- Child-Pugh classification when available 5
The combination of K70.31 with documentation of diuretic use (loop and/or potassium-sparing diuretics) provides the highest accuracy for identifying these patients in administrative databases, with sensitivity and specificity exceeding 82% and 89% respectively 1.