GastroAGI Logo
OverviewBlogsAbout
Trending TopicsDaily BriefConference

116 calculators match

Most used

21
MELD-NaAssesses the severity of chronic liver diseaseChild-Pugh ScoreAssesses the prognosis of chronic liver disease, mainly cirrhosisFIB-4 IndexLiver fibrosis scoring indexAPRIAST to platelet ratio — liver fibrosisMaddrey's DFAlcoholic hepatitis severityGlasgow-BlatchfordUpper GI bleed risk stratificationGAHSGlasgow alcoholic hepatitis scoreMontreal IBDIBD classification — CD & UCMayo ScoreUlcerative colitis activityBISAP ScoreBedside index for severity of pancreatitisCLIF-SOFAOrgan failure scoring in cirrhosisAlcohol ContentStandard drinks & alcohol grams calculatorAARC-ACLFAcute-on-chronic liver failure gradePELD / CR ScorePediatric end-stage liver diseaseHarvey-BradshawCrohn's disease activity indexCTSICT severity index — pancreatitisRockall ScoreGI bleed rebleeding & mortality riskCAGEAlcohol use disorder screening (4 questions)MELD 3.0Updated MELD — sex-inclusive formulaAUDIT ScoreAlcohol use disorders identification testVOCAL-Penn ScorePost-operative mortality risk in cirrhosis surgery

Liver & Cirrhosis

17
ALBI GradeAlbumin-bilirubin liver function grade in HCCUKELD ScoreUK model for end-stage liver diseaseMELD-XIMELD excluding INR — for anticoagulated patientsWest Haven CriteriaHepatic encephalopathy gradingMilan CriteriaLiver transplant eligibility in hepatocellular carcinomaLI-RADS v2018 (CT/MRI)Liver observation category from size, APHE and major featuresBCLC StagingHepatocellular carcinoma stage and treatment allocationSimplified AIH CriteriaSimplified criteria for autoimmune hepatitisRevised Original AIH ScoreIAIHG 1999 comprehensive autoimmune hepatitis scoreSAAGSerum-ascites albumin gradient — cause of ascitesR FactorHepatocellular vs cholestatic pattern in liver injuryCLIF-C ACLFMortality prediction in acute-on-chronic liver failureKing's College CriteriaTransplant criteria in acute liver failureGALAD ScoreHCC detection from gender, age, AFP-L3, AFP and DCPMetroticket 2.0AFP-adjusted up-to-seven for HCC transplant eligibilityRUCAMCausality in drug- and herb-induced liver injuryBaveno VII CriteriacACLD, CSPH and sparing screening endoscopy

Fibrosis & MASLD

8
NAFLD Fibrosis ScoreAdvanced fibrosis probability in MASLD/NAFLDBARD ScoreBMI, AST/ALT ratio, diabetes — MASLD fibrosisFatty Liver IndexPredicts hepatic steatosis from routine labsFibrotic NASH Index (FNI)At-risk NASH probability from AST, HbA1c and HDLNAFLD Activity Score (NAS)Histologic activity grade — steatosis, inflammation, ballooningMEFIB IndexMRE + FIB-4 rule for significant fibrosis (≥F2) in MASLDFAST ScoreFibroScan-AST — at-risk NASH from LSM, CAP and ASTSAFE ScoreSteatosis-Associated Fibrosis Estimator for MASLD in primary care

Pancreas & Biliary

8
Ranson's CriteriaAcute pancreatitis severity at 48 hoursGlasgow-Imrie CriteriaAcute pancreatitis severity — the PANCREAS criteriaHAPSHarmless acute pancreatitis scoreTokyo Guidelines — CholangitisTG18 diagnosis and severity grade for acute cholangitisTokyo Guidelines — CholecystitisTG18 diagnosis and severity grade for acute cholecystitisBiliary Pain (Rome IV)Rome IV — defining biliary-type pain before interventionFunctional Pancreatic SODRome IV — pancreatic sphincter of Oddi disorderRevised Atlanta ClassificationAcute pancreatitis severity — mild, moderately severe, severe

IBD

9
Truelove & Witts CriteriaAcute severe ulcerative colitis — admission decisionUCEISUlcerative colitis endoscopic index of severitySCCAISimple clinical colitis activity index — symptoms onlyCDAICrohn's disease activity index — the trial standardSES-CDEndoscopic severity in Crohn's diseasePUCAIPaediatric ulcerative colitis activity indexTravis (Oxford) CriteriaDay 3 colectomy risk in acute severe ulcerative colitisHo IndexDay 3 steroid failure risk in acute severe ulcerative colitisRutgeerts ScorePostoperative Crohn's recurrence at ileocolonoscopy

GI Bleeding

7
EVendo ScorePredicts oesophageal varices needing treatmentForrest ClassificationPeptic ulcer bleeding — rebleeding risk at endoscopyAIMS65 ScoreUpper GI bleed mortality — five bedside criteriaOakland ScoreSafe-discharge risk for acute lower GI bleedingABC ScoreAge, blood tests, comorbidities — GI bleed mortalitySarin ClassificationEndoscopic classification of gastric varicesEGUS (Gastric Ulcer)Malignancy risk in a gastric ulcer, and who needs repeat endoscopy

Alcohol

2
ABIC ScoreAge, bilirubin, INR, creatinine — alcoholic hepatitisLille ModelSteroid response at day 7 in alcoholic hepatitis

Upper GI

4
Chicago Classification v4.0Oesophageal motility pattern from high-resolution manometryLA Classification (Oesophagitis)Los Angeles grade A–D for erosive oesophagitisPrague C & M CriteriaCircumferential and maximal extent of Barrett's oesophagusEREFS (Eosinophilic Oesophagitis)Endoscopic reference score — oedema, rings, exudates, furrows, stricture

Colorectal

3
Boston Bowel Prep ScaleColonoscopy preparation adequacy by segmentStool Osmotic GapOsmotic vs secretory diarrhoea from stool electrolytesATLAS Score (C. difficile)Predicted response to therapy in Clostridioides difficile infection

Functional GI

37
Bristol Stool ScaleStool form types 1–7 and colonic transitRome IV Criteria for IBSIrritable bowel syndrome diagnosis and subtypeFunctional ConstipationRome IV — two of six items, IBS excludedOpioid-Induced ConstipationRome IV — constipation tied to opioid therapyFunctional DiarrhoeaRome IV — loose stools without predominant painFunctional Bloating / DistensionRome IV — bloating without other bowel disorder criteriaUnspecified Functional Bowel DisorderRome IV — bowel symptoms fitting no other categoryCentrally Mediated Abdominal Pain (CAPS)Rome IV — continuous pain unrelated to gut eventsNarcotic Bowel SyndromeRome IV — opioid-induced hyperalgesia of the gutFaecal Incontinence (Rome IV)Rome IV — the criteria, and why nobody is askedFunctional Anorectal PainLevator ani, unspecified pain and proctalgia fugaxFunctional Defecation DisordersRome IV — dyssynergia and inadequate propulsionInfant RegurgitationRome IV — the happy spitter, and the alarm features that rule it outInfant ColicRome IV — recurrent unexplained crying in a well infant under 5 monthsInfant DyscheziaRome IV — straining before a soft stool, and why not to intervenePaediatric Functional ConstipationRome IV — two of six over one month, with overflow soiling as a criterionToddler's DiarrhoeaRome IV functional diarrhoea of childhood — painless, thriving childPaediatric Cyclic Vomiting SyndromeRome IV — both age bands, with different criteria for eachPaediatric Rumination SyndromeRome IV — infant and child/adolescent criteriaFunctional Nausea & Vomiting (Children)Rome IV — two separate disorders that can be met togetherAerophagiaRome IV — distension that increases through the dayPaediatric Functional DyspepsiaRome IV — four times a month, with PDS and EPS subtypingPaediatric Irritable Bowel SyndromeRome IV — plus the constipation clause clinicians missAbdominal MigraineRome IV — stereotypical incapacitating episodes weeks apartFunctional Abdominal Pain — NOSRome IV — the residual category, reached after the other threeNonretentive Faecal IncontinenceRome IV — soiling without retention, where laxatives make it worseFunctional DyspepsiaRome IV — with PDS and EPS subtypingRumination SyndromeRome IV — effortless regurgitation without retchingCyclic Vomiting SyndromeRome IV — stereotypical episodic vomitingCannabinoid HyperemesisRome IV — CVS pattern relieved by cannabis cessationChronic Nausea & VomitingRome IV — chronic nausea and vomiting syndromeBelching DisordersRome IV — supragastric vs gastric belchingFunctional HeartburnRome IV — heartburn with normal acid exposureReflux HypersensitivityRome IV — normal acid exposure, positive symptom associationFunctional Chest PainRome IV — non-cardiac, non-reflux chest painGlobusRome IV — painless lump-in-throat sensationFunctional DysphagiaRome IV — dysphagia with normal endoscopy and manometry
  1. Calculators
  2. /
  3. SAAG
Liver & Cirrhosis

SAAG

Serum-ascites albumin gradient — cause of ascites

Both samples must be taken on the same day — ideally the same hour — because the gradient is a difference between two moving values. It answers one question only: is this ascites caused by portal hypertension?

When to use
Use it on the first diagnostic paracentesis in any patient with new ascites, and again whenever the cause is in doubt or a second process is suspected. It answers one question — is this ascites driven by portal hypertension? — and it answers it well enough that it should be part of the initial fluid panel alongside cell count, culture and total protein. It does not need repeating on every therapeutic tap once the cause is established, because the gradient is a property of the mechanism rather than of the current fluid volume.
Why use it
Because the question that actually changes management is mechanistic, not descriptive. The older exudate-transudate framework, borrowed from pleural fluid, classified ascites by total protein concentration and misclassified a substantial proportion of patients — cirrhotic ascites with a high protein content was labelled exudative and sent down the wrong diagnostic path. Runyon's group showed that the albumin gradient reflects the underlying oncotic-hydrostatic balance in the sinusoids directly, and therefore tracks portal pressure rather than the incidental protein content of the fluid. Its roughly 97% accuracy is high enough that a confidently high or low gradient reshapes the differential immediately, at the cost of one extra albumin request on a sample already being sent.
Formula, evidence and interpretation

About the Serum-Ascites Albumin Gradient (SAAG)

Subtract the ascitic fluid albumin from the serum albumin, both drawn the same day. A gradient of 1.1 g/dL or above indicates portal hypertension with about 97% accuracy; below 1.1 g/dL points away from it, towards peritoneal disease such as carcinomatosis, tuberculosis or pancreatic ascites. The gradient identifies a mechanism rather than a disease — high-SAAG ascites includes cardiac failure and Budd-Chiari as well as cirrhosis, which is why the ascitic total protein is the necessary second test. SAAG replaced the exudate-transudate classification because that older approach misclassified a substantial minority of patients.

On this page

  • Formula
  • Interpreting the result
  • Inputs
  • What it returns
  • How it is calculated
  • Facts & figures
  • Evidence
  • How it compares
  • Pearls & pitfalls
  • Critical actions
  • Why it exists
  • About the creator
  • Limitations
  • If you are the patient
  • FAQ
  • Related calculators
  • References

Formula

SAAG = serum albumin (g/dL) − ascitic fluid albumin (g/dL)
serum albumin
Measured on the same day as the ascitic sample.
ascitic fluid albumin
Measured on the diagnostic paracentesis sample. Not interchangeable with ascitic total protein, which is a separate and separately useful test.
  • It is a difference, not a ratio. There is no division and no reference range applied to either value individually.
  • The single cut-off is 1.1 g/dL (11 g/L). A gradient of exactly 1.1 counts as portal hypertension.
  • Both samples must be same-day. Albumin infusion, large-volume paracentesis and aggressive diuresis all move the serum value, and a gradient assembled from values days apart is uninterpretable.
  • A low serum albumin does not invalidate the gradient. Severe hypoalbuminaemia lowers both values roughly in parallel, so the difference is preserved — this is precisely why a gradient outperforms either concentration alone.
  • Ascitic total protein is a different test answering a different question, and should be requested alongside rather than instead.

Interpreting the result

Read the result as a mechanism and then immediately ask which disease within that mechanism. A gradient of 1.1 g/dL or above means portal hypertension — most often cirrhosis, but also cardiac ascites, Budd-Chiari syndrome, portal vein thrombosis, sinusoidal obstruction syndrome and massive hepatic metastases. Distinguishing cirrhotic from cardiac ascites is the common practical problem within this band, and the ascitic total protein does it: cardiac ascites is typically 2.5 g/dL or above while cirrhotic ascites is usually below that. A gradient under 1.1 points to peritoneal disease and redirects the workup to cytology, adenosine deaminase and mycobacterial culture, and fluid amylase. One important caveat sits underneath both bands: mixed ascites exists. A cirrhotic patient who develops peritoneal carcinomatosis or tuberculous peritonitis may produce an intermediate or discordant gradient, so a result that conflicts with the clinical picture deserves a second look rather than deference.

ScoreBandWhat it meansAction
≥ 1.1 g/dLPortal hypertensionApproximately 97% accurate for portal hypertension. Covers cirrhosis, cardiac ascites, Budd-Chiari, portal vein thrombosis, sinusoidal obstruction syndrome and massive liver metastasesSodium restriction and diuretics; check ascitic total protein to separate cardiac from cirrhotic, and cell count to exclude SBP
< 1.1 g/dLNot portal hypertensionPoints to peritoneal disease: carcinomatosis, tuberculous peritonitis, pancreatic ascites, nephrotic syndrome, serositisSend cytology (and repeat it), adenosine deaminase and mycobacterial culture, and fluid amylase as the picture dictates

What the SAAG needs (2 inputs)

Serum albumin
Entered in g/dL with a g/L toggle. Must be from the same day as the ascitic sample — ideally the same hour, since albumin shifts with infusions, diuresis and paracentesis.
Ascitic fluid albumin
From the diagnostic tap. Request it explicitly: many laboratories run ascitic total protein by default and albumin only if asked, and total protein cannot substitute.

Units. Albumin is entered in g/dL with a g/L toggle. The cut-off of 1.1 g/dL is 11 g/L. Because the calculation is a subtraction rather than a ratio, both values must be in the same unit — the toggle handles that, but if converting by hand, divide g/L by 10. Note that this calculator rounds the gradient to two decimal places before comparing it with the cut-off, which matters: in binary floating point some clinically ordinary pairs (3.3 − 2.2, for instance) evaluate marginally below 1.1 and would otherwise be misclassified.

What it returns

Albumin gradient
The arithmetic difference in g/dL. Reported to two decimal places; the only cut-off is 1.1.
Mechanism
Portal hypertension, or not portal hypertension. Deliberately not a disease label — several diseases sit on each side.

How it is calculated

The gradient is a bedside readout of Starling forces across the hepatic sinusoid. In portal hypertension the hydrostatic pressure in the sinusoids is high, so fluid is forced out while albumin is comparatively retained in the vascular compartment, producing ascites that is relatively albumin-poor and therefore a wide gradient. In peritoneal disease the leak comes from inflamed or infiltrated peritoneal surfaces and capillaries whose permeability is increased, so albumin escapes with the fluid and the ascitic albumin approaches the serum value, narrowing the gradient. Because both terms move together when the patient is globally hypoalbuminaemic, the difference remains informative in exactly the malnourished cirrhotic population where an absolute ascitic protein or albumin concentration would be misleading. That robustness is what the 1992 study demonstrated against the exudate-transudate concept, and why the older classification was abandoned.

Facts & figures

Causes by gradient
SAAG ≥ 1.1 (portal hypertension)SAAG < 1.1 (not portal hypertension)
CirrhosisPeritoneal carcinomatosis
Alcohol-related hepatitisTuberculous peritonitis
Cardiac ascitesPancreatic ascites
Budd-Chiari syndromeNephrotic syndrome
Portal vein thrombosisSerositis in connective tissue disease
Sinusoidal obstruction syndromeBowel obstruction or infarction
Massive hepatic metastasesBiliary ascites

Mixed ascites — most often cirrhosis with superimposed peritoneal carcinomatosis or tuberculosis — can produce a gradient that fits neither column cleanly.

The second test: ascitic total protein within a high gradient
Ascitic total proteinSuggestsWhy it matters
≥ 2.5 g/dLCardiac ascites (or Budd-Chiari, early)Sinusoids are not yet capillarised, so protein still escapes — treatment is directed at the heart, not the liver
< 2.5 g/dLCirrhotic ascitesCapillarised sinusoids retain protein
< 1.5 g/dLCirrhotic, and at higher risk of SBPIdentifies patients who may benefit from primary antibiotic prophylaxis

This is why total protein should be requested alongside albumin rather than instead of it — the two answer different questions and are both cheap.

Evidence

Validation against the exudate-transudate concept — Runyon et al.

1992

Prospective comparison of the serum-ascites albumin gradient against the exudate-transudate classification in patients undergoing diagnostic paracentesis, with the cause of ascites established independently.

The albumin gradient classified ascites by the presence or absence of portal hypertension with approximately 97% accuracy, substantially outperforming the exudate-transudate concept, which misclassified a considerable proportion of patients — particularly cirrhotic ascites with a high protein content.

Original description — Paré et al.

1983

The earlier study describing the serum-ascites albumin concentration gradient as a physiological approach to the differential diagnosis of ascites, on which the 1992 validation built.

Established the physiological rationale linking the gradient to sinusoidal Starling forces and portal pressure, rather than to the incidental protein content of the fluid.

AASLD practice guidance

2021

The 2021 AASLD guidance on the diagnosis, evaluation and management of ascites, spontaneous bacterial peritonitis and hepatorenal syndrome.

Recommends the albumin gradient as the initial classification step on diagnostic paracentesis, alongside cell count with differential, culture and total protein.

How it compares

SAAG vs Exudate-transudate classification by ascitic total protein

Superseded. SAAG classifies ascites by portal hypertension with about 97% accuracy, while the exudate-transudate concept misclassified a substantial minority.

The older framework used ascitic total protein above or below 2.5 g/dL, borrowed from pleural fluid analysis. It fails in ascites because the protein concentration depends on the patient's own serum albumin and on the degree of sinusoidal capillarisation, so cirrhotic ascites with a high protein content was routinely misclassified as exudative. Taking a difference rather than an absolute concentration removes that dependence. Total protein remains useful, but for a different question — separating cardiac from cirrhotic ascites within a high gradient.

Runyon BA, Montano AA, Akriviadis EA, Antillon MR, Irving MA, McHutchison JG. The serum-ascites albumin gradient is superior to the exudate-transudate concept in the differential diagnosis of ascites. Ann Intern Med. 1992;117(3):215-220.

SAAG vs Child-Pugh score

Different questions — SAAG asks why there is ascites, Child-Pugh grades how advanced the liver disease is once cirrhosis is established.

Child-Pugh includes ascites as one of its five components, so it presupposes that the ascites has already been attributed. SAAG is the test that does the attributing. In practice SAAG comes first on the diagnostic tap, and Child-Pugh follows once cirrhosis is the established cause.

Open the Child-Pugh score calculator →

SAAG vs MELD-Na

Complementary — SAAG is diagnostic, MELD-Na is prognostic and allocative, and neither substitutes for the other.

Once a high gradient has established portal-hypertensive ascites from cirrhosis, MELD-Na quantifies severity and transplant priority. SAAG has no prognostic content at all; a gradient of 2.5 does not indicate worse disease than one of 1.5.

Open the MELD-Na calculator →

Pearls & pitfalls

  • Request ascitic ALBUMIN explicitly. Many laboratories run total protein by default, and total protein cannot be substituted into this calculation.
  • Both samples must be same-day. Albumin infusion, large-volume paracentesis and brisk diuresis all move the serum value within hours.
  • A low serum albumin does not invalidate the result. Both terms fall together, so the difference is preserved — that robustness is the whole point of using a gradient.
  • The gradient gives a mechanism, not a diagnosis. A high gradient in a patient with no liver disease should prompt an echocardiogram before it prompts a liver biopsy.
  • Always send total protein as well. Within a high gradient, 2.5 g/dL or above suggests cardiac ascites and below 1.5 g/dL flags SBP prophylaxis candidates.
  • Mixed ascites exists. A cirrhotic who develops peritoneal carcinomatosis or tuberculosis can produce a discordant gradient, so trust a strongly conflicting clinical picture over the number.
  • Send cell count and culture from the same tap regardless of the gradient — 250 neutrophils/mm³ or above defines SBP and that decision is independent of the cause of the ascites.
  • A single negative cytology does not exclude carcinomatosis. It needs repeating, ideally on a large-volume sample.

Critical actions

  • Send albumin, total protein, cell count with differential, and culture in blood culture bottles at the bedside from every diagnostic tap.
  • Draw the serum albumin the same day as the paracentesis.
  • Where the gradient is high but there is no known liver disease, arrange an echocardiogram — cardiac ascites is the commonly missed cause in this band.
  • Where the gradient is low, send cytology and repeat it, plus adenosine deaminase and mycobacterial culture where tuberculosis is plausible.
  • Do not defer the diagnostic tap in a patient with ascites and any deterioration; SBP is frequently asymptomatic and the tap answers that question too.
  • Reconsider the result rather than accept it when it conflicts strongly with the clinical picture — mixed ascites is the usual explanation.

Why this score exists

The important conceptual move was to stop borrowing from pleural fluid. The exudate-transudate framework classifies effusions by how much protein they contain, which works reasonably in the pleural space but fails in the abdomen because ascitic protein content depends heavily on the patient's serum albumin and on how capillarised the hepatic sinusoids have become. A malnourished cirrhotic and a well-nourished one produce different ascitic protein concentrations from identical portal pressures. Taking the difference between serum and ascitic albumin cancels that shared term out, which is why a gradient is robust where a concentration is not — and why the older classification, which had been applied to ascites for decades essentially by analogy, turned out to be measurably worse.

About the creator

  • Bruce A. Runyon

    First author, 1992 validation study

    Led the study establishing the albumin gradient as superior to the exudate-transudate concept, and much of the subsequent guidance on ascites management.

  • Pierre Paré

    First author of the 1983 paper originally describing the serum-ascites albumin concentration gradient.

Limitations

  • Identifies a mechanism, not a disease. Everything from cirrhosis to constrictive pericarditis sits in the high-gradient band and needs separating by other means.
  • Loses reliability in mixed ascites, where portal hypertension and peritoneal disease coexist — exactly the situation in which the answer would be most useful.
  • Depends on same-day sampling; albumin infusion, large-volume paracentesis and diuresis all shift the serum term within hours.
  • Reported accuracy of roughly 97% comes from a population in which the cause was ultimately established, which will overstate performance in genuinely undifferentiated cases.
  • Chylous ascites and heavily bloody taps can distort the albumin measurement itself.
  • Says nothing about infection. A high or low gradient has no bearing on whether spontaneous bacterial peritonitis is present, which is decided by the neutrophil count.
  • No prognostic content — the size of the gradient above 1.1 carries no additional information.

If you are the patient

Ascites is a build-up of fluid in the abdomen, and it has quite different causes that need quite different treatments. To work out which, a doctor takes a small sample of the fluid with a needle and measures the albumin — a protein — in both the fluid and a blood sample taken the same day. Subtracting one from the other gives the SAAG. If the difference is 1.1 or more, the fluid is being driven by high pressure in the veins around the liver, most commonly from cirrhosis but sometimes from a heart problem; treatment then focuses on salt restriction and water tablets. If the difference is less than 1.1, the cause is more likely something affecting the lining of the abdomen itself, such as infection or cancer, and further tests on the fluid follow. The test is about 97% accurate at telling those two situations apart, which is why it is done on almost every first fluid sample.

Frequently asked questions

What is the SAAG?#

The serum-ascites albumin gradient: serum albumin minus ascitic fluid albumin, both measured the same day. A gradient of 1.1 g/dL or above indicates portal hypertension with roughly 97% accuracy.

What does a SAAG of 1.1 or above mean?#

Portal hypertension. Most often cirrhosis, but the band also includes cardiac ascites, Budd-Chiari syndrome, portal vein thrombosis, sinusoidal obstruction syndrome and massive hepatic metastases. It identifies the mechanism, not the disease.

What does a SAAG below 1.1 mean?#

That portal hypertension is unlikely, and the cause is probably peritoneal — carcinomatosis, tuberculous peritonitis, pancreatic ascites, nephrotic syndrome or serositis. Cytology, adenosine deaminase and fluid amylase become the relevant tests.

Can ascitic total protein be used instead of albumin?#

No. The gradient is defined on albumin, and total protein answers a different question. Within a high gradient, an ascitic total protein of 2.5 g/dL or above suggests cardiac rather than cirrhotic ascites, and below 1.5 g/dL identifies patients who may benefit from SBP prophylaxis — so send both.

Does a low serum albumin make the SAAG unreliable?#

No, and this is the test's main advantage. Severe hypoalbuminaemia lowers the serum and ascitic values roughly in parallel, so the difference between them is preserved. That is exactly why a gradient outperforms an absolute ascitic protein concentration.

How do I tell cardiac ascites from cirrhotic ascites?#

Both give a high gradient, so use the ascitic total protein. Cardiac ascites is typically 2.5 g/dL or above because the sinusoids are not capillarised and protein still escapes; cirrhotic ascites is usually below that. A high gradient in a patient with no known liver disease should prompt an echocardiogram.

Why did SAAG replace the exudate-transudate classification?#

Because the older approach, borrowed from pleural fluid, classified ascites by total protein concentration and misclassified a substantial minority — particularly cirrhotic ascites with a high protein content. Runyon's 1992 study showed the albumin gradient was markedly more accurate.

Does the SAAG need repeating on every paracentesis?#

No. Once the cause is established, the gradient reflects an unchanged mechanism and adds nothing. Cell count and culture, however, should be sent on any tap where infection is a possibility.

Related calculators

  • Child-Pugh Score — Assesses the prognosis of chronic liver disease, mainly cirrhosis
  • MELD-Na — Assesses the severity of chronic liver disease
  • CLIF-SOFA — Organ failure scoring in cirrhosis
  • Baveno VII Criteria — cACLD, CSPH and sparing screening endoscopy
  • ALBI Grade — Albumin-bilirubin liver function grade in HCC
  • UKELD Score — UK model for end-stage liver disease

References

Original / primary reference

  1. Runyon BA, Montano AA, Akriviadis EA, Antillon MR, Irving MA, McHutchison JG. The serum-ascites albumin gradient is superior to the exudate-transudate concept in the differential diagnosis of ascites. Ann Intern Med. 1992;117(3):215-220.
  2. Paré P, Talbot J, Hoefs JC. Serum-ascites albumin concentration gradient: a physiologic approach to the differential diagnosis of ascites. Gastroenterology. 1983;85(2):240-244.

Guidelines

  1. Biggins SW, Angeli P, Garcia-Tsao G, et al. Diagnosis, Evaluation, and Management of Ascites, Spontaneous Bacterial Peritonitis and Hepatorenal Syndrome: 2021 Practice Guidance by the American Association for the Study of Liver Diseases. Hepatology. 2021;74(2):1014-1048.
  2. European Association for the Study of the Liver. EASL Clinical Practice Guidelines for the management of patients with decompensated cirrhosis. J Hepatol. 2018;69(2):406-460.

Last updated August 1, 2026. Clinical knowledge base written and curated by GastroAGI Team from primary medical literature.

Written from primary literature and not yet independently clinically reviewed.

For use by qualified healthcare professionals. This calculator supports clinical judgement and does not replace it.