AIMS65 Score Calculator
AIMS65 Score Calculator
Five equally weighted items — albumin, INR, mental status, systolic pressure and age — scored for in-hospital mortality in upper gastrointestinal bleeding. The albumin threshold is 3.0 g/dL, not 3.0 mg/dL.
AIMS65 score
5 items → 0–5Albumin 2.8 g/dL, INR 1.8, no altered mental status, systolic 105 mmHg, age 72
Scoring
Range 0–5, one point each, no weighting
- g/dL, not mg/dL
- the albumin threshold is 3.0 g/dL, which is 30 g/L. Both full texts this page’s item list was read in print “mg/dL”, as does much of the secondary literature — a typographical error copied repeatedly. No serum albumin is 3 mg/dL
- one inclusive threshold and two exclusive
- systolic pressure is ≤ 90, so exactly 90 scores. Albumin is strictly < 3.0 and age strictly > 65, so exactly 3.0 and exactly 65 do not. Mixing these up moves a small number of patients a whole point, and nothing in the output shows it has happened
- unweighted, deliberately
- all five items score 1, which is why it can be computed at a bedside without a table — and why it cannot express that one item carries more information than another
- mortality, not intervention
- derived against in-hospital death, length of stay and cost. Head-to-head studies find Glasgow-Blatchford better for transfusion and endoscopic therapy and AIMS65 better for death
- the albumin is doing a lot of work
- albumin falls as an acute-phase response in any serious illness and is diluted by resuscitation fluid, so it is partly severity and partly timing. A sample drawn after several litres of crystalloid is not the sample the score was derived on
Worked example
Albumin 2.8 g/dL, INR 1.8, no altered mental status, systolic 105 mmHg, age 72
Albumin 2.8 is below 3.0 g/dL — 1
INR 1.8 is above 1.5 — 1
Mental status normal — 0 · systolic 105 is above 90 — 0 · age 72 is over 65 — 1
Total 3 points. In-hospital mortality at 3 was 1.5% in Park's cohort of 634 patients
Enter the same albumin as 28 g/L with the unit set to g/L and the answer is the same 3. Enter 28 with the unit left on g/dL and the albumin item scores 0 — the score becomes 2, which is the error the unit selector exists to prevent
Set the albumin to exactly 3.0 g/dL and the item scores 0, because the comparison is strictly below. Set the systolic to exactly 90 and that item scores 1, because that comparison is inclusive
All five items negative gives 0; all five positive gives 5, the maximum
The five AIMS65 items
| Letter | Item | Threshold | Points |
|---|---|---|---|
| A | Albumin | Below 3.0 g/dL (30 g/L) — strictly below | 1 |
| I | INR | Above 1.5 — strictly above | 1 |
| M | Mental status | Altered | 1 |
| S | Systolic blood pressure | 90 mmHg or below — inclusive | 1 |
| 65 | Age | Over 65 years — strictly over | 1 |
In-hospital mortality by AIMS65 score
| AIMS65 | In-hospital mortality | Cohort |
|---|---|---|
| 0 | 0% | 634 patients, non-variceal UGIB, Korea |
| 1 | 0% | same cohort |
| 2 | 0.9% | same cohort |
| 3 | 1.5% | same cohort |
| 4 | 9.5% | same cohort |
| 5 | 50.0% | same cohort |
Five items, one point each, and a mortality endpoint
AIMS65 was derived from a large administrative dataset to predict in-hospital mortality, length of stay and cost in acute upper gastrointestinal bleeding, and its appeal is that it can be computed at a bedside without a table: albumin below 3.0 g/dL, INR above 1.5, altered Mental status, Systolic pressure of 90 mmHg or below, and age over 65, one point each. The acronym does the remembering.
What it predicts matters, because the three upper gastrointestinal bleeding scores in common use are not interchangeable. AIMS65 was built against death. The Glasgow-Blatchford score was built against the need for hospital-based intervention. The Rockall score was built against mortality too, but with endoscopic items that make its full version unavailable at presentation. Head-to-head studies find roughly what that design history predicts: Glasgow-Blatchford better for transfusion and endoscopic therapy, AIMS65 better for death. In Park’s cohort of 634 patients AIMS65 reached an area under the curve of 0.943 for in-hospital mortality against 0.856 for the Rockall score, and the systematic review quoted in the ESGE 2021 guideline gives it a pooled sensitivity of 0.79 and a specificity of 0.61 for 30-day serious adverse events — a different endpoint, and a much less flattering figure.
Two practical points. The albumin threshold is in grams per decilitre. Both of the full texts in which this page’s item list was checked print it as “albumin < 3.0 mg/dL”, and that error is reproduced widely; the value is 3.0 g/dL, which is 30 g/L, and no serum albumin is three milligrams per decilitre. The unit selector above exists for that. And the three numeric thresholds do not share an edge convention: the systolic pressure item includes exactly 90 while the albumin and age items exclude exactly 3.0 and exactly 65.
The last thing to know is what albumin is doing in a mortality score for bleeding. It falls as an acute-phase response in any serious illness and it is diluted by resuscitation fluid, so it is partly a measure of how ill the patient is and partly an artefact of when the sample was taken. An albumin drawn after several litres of crystalloid is not the measurement the score was derived on. A score is not a diagnosis and a risk derived from a cohort is not a probability for one patient. This states what the number predicted in a named study; the clinician in front of the patient decides what follows. Thresholds and reference intervals are method- and laboratory-dependent, and your own laboratory’s interval takes precedence over any figure quoted here.
Frequently asked questions
What does AIMS65 stand for?
Albumin below 3.0 g/dL, INR above 1.5, altered Mental status, Systolic blood pressure 90 mmHg or below, and age over 65 years. One point each, maximum 5, with no weighting.
Is the AIMS65 albumin threshold 3.0 g/dL or 3.0 mg/dL?
3.0 g/dL, which is 30 g/L. The “mg/dL” that appears in much of the published literature — including both of the full texts this page’s items were checked against — is a typographical error copied from paper to paper. A serum albumin of 3 mg/dL does not occur.
Does a systolic pressure of exactly 90 score a point?
Yes — that item is “90 mmHg or below”. The albumin and age items are the other way round: strictly below 3.0 g/dL and strictly over 65 years, so exactly 3.0 and exactly 65 score nothing. The thresholds do not share an edge convention.
What does an AIMS65 score of 0 mean?
In Park’s cohort of 634 patients with non-variceal upper gastrointestinal bleeding, in-hospital mortality was 0% at scores of 0 and 1. At that cohort size a zero event rate is still consistent with a true rate of around 1%, and the score says nothing about the need for transfusion or endoscopic therapy.
AIMS65 or Glasgow-Blatchford?
They were derived against different endpoints and the evidence follows the design. AIMS65 discriminates mortality better; the Glasgow-Blatchford score discriminates the need for transfusion and endoscopic intervention better, and the ESGE 2021 guideline recommends it for pre-endoscopy risk stratification. The Rockall score needs the endoscopy findings for its full form.
Related calculators
References
- Saltzman JR, Tabak YP, Hyett BH, Sun X, Travis AC, Johannes RS. A simple risk score accurately predicts in-hospital mortality, length of stay, and cost in acute upper GI bleeding. Gastrointest Endosc. 2011;74(6):1215–24.
- Park SM, Yeum SC, Kim BW, et al. Comparison of AIMS65 score and other scoring systems for predicting clinical outcomes in Koreans with nonvariceal upper gastrointestinal bleeding. Clin Endosc. 2015;48(5):380–7.
- Kim MS, Choi J, Shin WC. AIMS65 scoring system is comparable to Glasgow–Blatchford score or Rockall score for prediction of clinical outcomes for non-variceal upper gastrointestinal bleeding. Clin Endosc. 2015;48(6):522–7.
- Gralnek IM, Stanley AJ, Morris AJ, et al. Endoscopic diagnosis and management of nonvariceal upper gastrointestinal hemorrhage (NVUGIH): European Society of Gastrointestinal Endoscopy (ESGE) Guideline — update 2021. Endoscopy. 2021;53(3):300–32.
Not medical advice. For healthcare professionals and education. Reference intervals vary by laboratory and assay — always use your own laboratory's. Never base a dose or a treatment decision on this page alone. Full disclaimer at calcengines.com/disclaimer/