Using MELD Score in Inpatient Management
The MELD score should be calculated on admission and serially monitored to predict short-term mortality risk, guide transplant referral decisions (particularly at MELD ≥15 or with any portal hypertensive complications), and stratify patients for appropriate level of care, while recognizing that MELD has significant limitations in capturing the full clinical picture of acutely decompensated patients. 1
Core Principles for Inpatient MELD Application
Initial Risk Stratification on Admission
Calculate the MELD score immediately upon hospital admission using objective laboratory values (bilirubin, creatinine, INR). MELD accurately stratifies patients by 90-day mortality risk without transplantation and serves as your baseline prognostic anchor 1.
The most recent MELD 3.0 iteration accounts for hyponatremia, female sex, and malnutrition—use this version rather than older calculations 1. For general inpatient populations with chronic liver disease, MELD scores of 15-19,20-29, and ≥30 show progressively increased mortality hazard ratios of 2.52,2.70, and 8.00 respectively 2.
Serial Monitoring and Dynamic Assessment
Do not rely on a single MELD value. The change in MELD over time (Delta MELD or ΔMELD) may be a better predictor of mortality than static scores 3. Specifically:
- Reassess MELD at 2 weeks: An increase of ≥2 points at 2 weeks is 75% sensitive and 76% specific for predicting 60-day mortality or need for transplantation 4
- If MELD does not increase by week 2, the patient has a 93.8% chance of 60-day survival 4
- Assessment at 1 week is too early (poor sensitivity/specificity), while 4 weeks is too late for meaningful intervention 4
Transplant Referral Thresholds
Refer patients for liver transplant evaluation at MELD ≥15 in the absence of portal hypertensive complications 1, 3. However, this is a guideline threshold, not an absolute barrier.
Critical caveat: Patients should be referred at ANY MELD score if they have uncontrolled portal hypertensive complications including:
- Refractory ascites
- Recurrent variceal bleeding
- Hepatic encephalopathy (recurrent or chronic)
- Hepatorenal syndrome (especially type I—expedited referral required) 3, 5
- Hepatopulmonary syndrome
- Portopulmonary hypertension
- Spontaneous bacterial peritonitis 5
These patients are "underrepresented by MELD"—their clinical severity exceeds what the score captures 1.
Recognizing MELD Limitations in Acute Settings
MELD significantly underestimates mortality in patients with acute-on-chronic liver failure (ACLF) because it captures intrinsic liver disease but not extrahepatic organ failures 6, 7. In ACLF patients:
- MELD and MELD-Na have lower discriminative ability (AUC ~0.70) compared to ACLF-specific scores like CLIF-C ACLF (AUC 0.79) 7
- Consider ACLF-specific scoring systems (CLIF-C ACLF, NACSELD ACLF, AARC score) for patients with multiorgan failure 6
- All patients with ACLF should be considered for urgent transplant center referral regardless of MELD score 1
Enhanced MELD Models for Inpatient Use
For hospitalized patients, particularly those critically ill, consider incorporating serum lactate to improve prognostic accuracy:
MELD-Lactate (MELD-LA) achieves superior discrimination (C-statistic 0.81-0.85) compared to MELD alone 8, 9. This is particularly valuable because:
- For MELD 25, inpatient mortality ranges from 11.2% (lactate 1 mmol/L) to >50% (lactate >8 mmol/L) 8
- MELD-LA improves risk prediction in 23.5% of patients compared to MELD alone 8
- Lactate clearance ≥35% within 24 hours predicts improved 30-day survival (OR 0.41-0.42) and may indicate treatment response 9
The MELD-LA formula is available at bswh.md/meldla for clinical use 8.
Level of Care Decisions
Use MELD score to guide intensity of monitoring and care:
MELD >30: These patients face extremely high mortality and morbidity risk. Address the risk-benefit of transplantation explicitly 3. Consider ICU-level care for:
- Need for organ support (vasopressors, mechanical ventilation, renal replacement therapy)
- Grade III-IV hepatic encephalopathy
- Massive bleeding or septic shock 7
MELD 15-29: Intermediate risk requiring close inpatient monitoring, aggressive management of complications, and prompt transplant evaluation if not already listed.
MELD <15: Generally lower short-term mortality risk, but do not use this as justification to delay referral if portal hypertensive complications are present 1.
Common Pitfalls to Avoid
Never use MELD score alone to deny transplant referral—clinical symptoms and complications trump the numerical score 1
Do not apply MELD to acute liver failure—all ALF patients require urgent transplant center referral regardless of calculated scores 1
Recognize MELD exceptions: HCC, hepatopulmonary syndrome, familial amyloidotic polyneuropathy, primary oxaluria, and other conditions require expert-assigned exception points 3
In patients with MELD <15, transplant may still show survival benefit—particularly with living donor transplantation (benefit reported at MELD as low as 11) 1
MELD does not predict post-transplant mortality except at very high scores (>35) 3
Practical Inpatient Algorithm
On admission:
- Calculate MELD 3.0 score
- Measure serum lactate if critically ill
- Assess for portal hypertensive complications
- Determine if patient meets ACLF criteria
During hospitalization:
- Monitor MELD at clinically relevant intervals (especially at 2 weeks)
- Track lactate clearance in septic/critically ill patients
- Escalate care if MELD increases ≥2 points
- Initiate transplant evaluation discussion if MELD ≥15 or complications present
At discharge/transition:
- Document MELD trajectory
- Ensure outpatient hepatology follow-up for MELD ≥10
- Expedite transplant referral for MELD ≥15 or any major complication
- Consider palliative care consultation for MELD >30 with poor trajectory 10, 11
The MELD score is a valuable objective tool but must be interpreted within the full clinical context, with particular attention to complications that the score does not capture and dynamic changes that signal deterioration or improvement.