ARISCAT Postoperative Pulmonary Risk Calculator
ARISCAT Postoperative Pulmonary Risk Calculator
Seven weighted predictors of pulmonary complications after surgery, from the Catalonian ARISCAT cohort. The points are unequal and two items alone are worth 24 — an intrathoracic incision, and an oxygen saturation of 90 per cent or below.
ARISCAT score
7 items → 0–12363 years old; SpO₂ 97 per cent on air; no recent respiratory infection; haemoglobin 128 g/L; upper abdominal incision; planned duration 2.5 hours; elective
Scoring
Under 26 low · 26 to 44 intermediate · 45 or more high
- which question this answers
- pulmonary complications — respiratory infection, respiratory failure, pleural effusion, atelectasis, pneumothorax, bronchospasm and aspiration pneumonitis — within the hospital stay. It says nothing about the heart, which is the RCRI‘s and the AUB-HAS2‘s question, and nothing about venous thromboembolism
- discrimination, and where it drifts
- Canet reported a c-index of 0.89 in development and 0.84 in validation for the simplified point score (0.90 and 0.88 for the full logistic model). External validation is less flattering: PERISCOPE found calibration drift outside the derivation setting, with the calibration slope falling from 0.81 in Western Europe to 0.58 in Eastern Europe, and a 2022 STARSurg systematic review reported that ARISCAT “failed to meet the a priori threshold for good discrimination (AUC >0.7)”. A Danish cohort found it underestimated complications in a high-risk population
- the unequal weights, which are the point
- two items are worth 24 — an intrathoracic incision and an SpO₂ of 90 per cent or below. A recent respiratory infection at 17 outweighs an upper abdominal incision at 15, and age over 80 at 16 outweighs both. The largest single step on the whole score is from 8 to 24 points as the saturation crosses from 91 to 90 per cent
- boundary conventions, printed because sources disagree
- the derivation article’s comparison symbols were garbled in every copy of it we could read, and independent reproductions render the anaemia cut as “10 g/dL or below” and as “under 10 g/dL”, and the lowest saturation band as “90 per cent or below” and as “under 91 per cent”. This page implements the inclusive reading in both cases. The difference is one patient at exactly 100 g/L and one at exactly 90 per cent, which is small and is not nothing
- which totals exist
- only 104 of the 124 integers from 0 to 123 can actually be made from these seven items, so a score of 21 is impossible and a score of 109 is impossible. Both sides of both class boundaries — 25 and 26, 44 and 45 — are reachable, so neither threshold is decorative
Worked example
63 years old; SpO₂ 97 per cent on air; no recent respiratory infection; haemoglobin 128 g/L; upper abdominal incision; planned duration 2.5 hours; elective
Age 3 + saturation 0 + infection 0 + anaemia 0 + incision 15 + duration 16 + emergency 0 = 34 points
34 sits in the 26 to 44 intermediate class
Canet's cohorts: 6.3 per cent in the development subsample and 13.3 per cent in validation. PERISCOPE observed 12.98 per cent in this class
Thirty-one of the 34 points came from the operation rather than from the patient. This is substantially a score about the incision and the clock
Drop the planned duration under two hours and the total falls to 18, low risk — a single theatre-scheduling assumption moves the class
Add a chest infection three weeks ago and the total rises to 51, high risk, where Canet's subsamples reported 44.9 and 42.1 per cent
That one item is worth more than the incision, and it is the item most often left unasked at a preoperative clinic
The seven predictors and their published points
| Predictor | Category | Points | Regression coefficient |
|---|---|---|---|
| Age | 51 to 80 years / over 80 | 3 / 16 | 0.331 / 1.619 |
| Preoperative SpO₂ | 91 to 95% / 90% or below | 8 / 24 | 0.802 / 2.375 |
| Respiratory infection in the last month | Yes | 17 | 1.698 |
| Preoperative anaemia | Haemoglobin 10 g/dL or below | 11 | 1.105 |
| Surgical incision | Upper abdominal / intrathoracic | 15 / 24 | 1.480 / 2.431 |
| Duration of surgery | 2 to 3 hours / over 3 hours | 16 / 23 | 1.593 / 2.268 |
| Emergency procedure | Yes | 8 | 0.768 |
What each class predicted, and in which cohort
| Class | Points | Canet development | Canet validation | PERISCOPE observed |
|---|---|---|---|---|
| Low | Under 26 | 0.7% | 1.6% | 3.39% (predicted 0.87%) |
| Intermediate | 26 to 44 | 6.3% | 13.3% | 12.98% (predicted 7.82%) |
| High | 45 or more | 44.9% | 42.1% | positive likelihood ratio 7.12 |
A score about the incision and the clock, more than about the patient
ARISCAT was derived from 2,464 patients having non-obstetric in-hospital surgery across 59 hospitals — every public-service hospital in Catalonia plus one centre in Valencia — recruited over a single year from January 2006. The overall rate of postoperative pulmonary complications was 5.0 per cent, and seven variables predicted it independently. Unlike most bedside scores the points are unequal, and the inequality is informative: an intrathoracic incision and a resting oxygen saturation of 90 per cent or below are each worth 24, while an emergency operation is worth 8.
Add up where the points come from and the character of the score becomes clear. The incision and the duration together can reach 47 of the 123 available, and a patient with nothing whatever wrong with them who is listed for a three-hour thoracotomy is already in the high-risk class. That is not a flaw — postoperative pulmonary complications genuinely are driven by where the knife goes and how long the lung is kept still — but it does mean that a high ARISCAT score is often a statement about the list rather than about the person on it, and that the two patient-side items a clinic can actually change are the saturation and the recent chest infection.
Of the patient-side items, the recent respiratory infection is the one worth asking about explicitly. At 17 points it outweighs an upper abdominal incision, and Canet’s definition is specific: an infection in the preceding month with fever and antibiotic treatment, not a current cough. The anaemia item, by contrast, is set at a haemoglobin of 10 g/dL — far below the level at which anaemia is usually acted on before surgery — so scoring zero there is not a statement that the haemoglobin is satisfactory.
The score’s discrimination was excellent in Catalonia and has been less so elsewhere. Canet reported a c-index of 0.89 in development and 0.84 in validation; the external PERISCOPE study found calibration drift outside the derivation setting, with the calibration slope falling from 0.81 in Western Europe to 0.58 in Eastern Europe, and observed roughly four times the predicted complication rate in the low-risk class. A 2022 systematic review concluded that ARISCAT failed to meet its threshold for good discrimination. It remains the most externally validated pulmonary risk model there is, which is a different claim from being accurate in any particular hospital. A score’s output is a cohort frequency, not this patient’s probability: a stratum in which 9 per cent had an event describes that stratum, not which 9 per cent. This page reports what a stratum predicted in a named study. It recommends no action.
Frequently asked questions
What does the ARISCAT score predict?
Pulmonary complications during the hospital stay after surgery — respiratory infection, respiratory failure, pleural effusion, atelectasis, pneumothorax, bronchospasm and aspiration pneumonitis. It says nothing about cardiac complications, which is what the Revised Cardiac Risk Index and AUB-HAS2 estimate, and nothing about venous thromboembolism.
Which ARISCAT items carry the most points?
Two items are worth 24: an intrathoracic incision, and a preoperative oxygen saturation of 90 per cent or below on room air. Age over 80 is worth 16 and a respiratory infection in the last month 17 — more than the 15 for an upper abdominal incision. The single biggest step on the score is from 8 to 24 points as the saturation crosses from 91 to 90 per cent.
Are the ARISCAT risk figures from the same cohort?
No, and they differ. Canet split his 2,464 patients into development and validation subsamples, which reported 0.7 against 1.6 per cent in the low class, 6.3 against 13.3 in the intermediate class, and 44.9 against 42.1 in the high class. Most secondary sources quote the validation figures without saying so. The external PERISCOPE cohort observed more complications than the model predicted in both lower classes.
Should the duration be the planned or the actual operating time?
Preoperatively it can only be the planned one, and that is a real weakness: the same patient scores differently before and after the same operation if the case overruns. Canet’s abstract describes the factor as a duration of at least two hours while the points table splits it at two and at three hours, so a two-hour case is the value to check against your own protocol.
Can every ARISCAT total between 0 and 123 occur?
No. Only 104 of the 124 integers are reachable from these seven items, so totals such as 21 and 109 cannot occur at all. Both sides of both class boundaries — 25 and 26, and 44 and 45 — are reachable, so neither threshold is decorative.
Related calculators
References
- Canet J, Gallart L, Gomar C, et al. Prediction of postoperative pulmonary complications in a population-based surgical cohort. Anesthesiology. 2010;113(6):1338–1350.
- Mazo V, Sabaté S, Canet J, et al. Prospective external validation of a predictive score for postoperative pulmonary complications. Anesthesiology. 2014;121(2):219–231.
- STARSurg Collaborative and EuroSurg Collaborative. Prognostic model research for postoperative pulmonary complications after major abdominal surgery: a systematic review. Br J Anaesth. 2022.
- MDCalc. ARISCAT score for postoperative pulmonary complications, which collects the PERISCOPE and STARSurg external-validation figures quoted here.
- Halvorsen S, Mehilli J, Cassese S, et al. 2022 ESC Guidelines on cardiovascular assessment and management of patients undergoing non-cardiac surgery. Eur Heart J. 2022;43(39):3826–3924.
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/