Risk Assessment in Anaesthesia and Critical Care: A Comprehensive Q&A
ASA Physical Status Classification System
What is the original purpose of the ASA Physical Status Classification System?
The history of the ASA system began around 1940 when the American Society of Anesthesiologists established a committee to study, examine, experiment, and devise a system for the collection and tabulation of statistical data in anesthesia.
It was developed as a system to collect and tabulate statistical data, not to do any risk prediction.
The original intent was never to use it as a risk prediction tool, but for statistical tabulation.
How has the ASA classification system evolved over time?
In 1941, the classification system was published with six designations, similar to today's but with differences like emergency surgery denoted as class five or six.
In 1961, Dripps and colleagues created a new system based on Saklad's original, which closely resembles the current one.
In 1962, the five-category system was adopted by the ASA House of Delegates.
In 1981, a modification changed ASA 5 wording. In 1983, an ASA 6 category was added for brain-dead organ donors.
In 2014, patient examples were added, and in 2020, pediatric and obstetric examples were included.
What are the current uses of the ASA classification in clinical practice?
Clinically, it helps anesthesia providers communicate patient status.
It is used for staff assignment and determining procedure location, for example, if a patient can be managed in an ambulatory surgery center versus a hospital.
It serves as an outcome predictor and has administrative use for comparing hospital performance.
It is integrated into billing algorithms as a modifier for some insurance companies.
In research, it helps describe the patient population in studies.
How well does the ASA physical status perform as a risk prediction tool?
While the ASA physical status is definitely associated with post-operative outcome, its positive predictive value for post-operative mortality is very low.
It is true that many ASA 4 or 5 patients have bad outcomes, but being an ASA 4 doesn't necessarily predict a bad outcome for an individual patient.
It performs poorly when used for individual patient risk prediction because it does not take procedure and care team or facility-related factors into account.
Complex patients must be condensed into a six-point scale, lacking the granularity needed for accurate individual risk prediction.
What factors make a good risk prediction system, and how does ASA compare?
A successful risk prediction system includes three large factors: patient-related factors (functional status, frailty, comorbidities), procedure-related factors (invasiveness, urgency), and team/facility-related factors (level of specialization, experience, equipment).
The ASA physical status classification system is designed to only look at the patient's physical factors at the time of pre-operative assessment.
Unlike validated tools like RCRI, NSQIP, or STS risk calculator, it does not include procedure type, operative severity, or emergency status, which are crucial for accurate risk prediction.
Risk Scoring for Non-Cardiac Surgery in Cardiac Patients
What are the major perioperative cardiac complications in non-cardiac surgery?
Major perioperative cardiac events complicate between 1.4% to 3.9% of surgeries.
Complications include myocardial infarction, pulmonary edema, ventricular fibrillation, cardiac arrest, or complete heart block.
The incidence is related to baseline risk and the timing of surgery after a recent myocardial event, with guidelines recommending at least a 60-day interval after acute coronary syndrome.
What is the Revised Cardiac Risk Index (RCRI) and how is it used?
The RCRI, published by Lee et al. in 1999, is a widely used risk index with six independent predictors: high-risk surgery, history of ischemic heart disease, history of congestive heart failure, history of cerebrovascular disease, pre-operative treatment with insulin, and pre-operative serum creatinine greater than 2 mg/dL.
Each risk factor is assigned one point. The risk of major cardiac event is less than 1% if none or one risk factor is present, 6.6% if two are present, and 11% if three or more are present.
What are the limitations of the Revised Cardiac Risk Index?
The RCRI poorly predicts post-operative mortality and does not accurately predict an individual's absolute risk of cardiac complications.
The component of diabetes mellitus may provide minimal prognostic information and could warrant elimination.
Renal dysfunction could be better defined using GFR, and surgical procedures would be better marked by level of operative complexity.
It underestimates risk in major vascular surgeries and doesn't include some low-risk procedures.
What is the Goldman Multifactorial Cardiac Risk Index?
The Goldman index, the original cardiac risk index, allocates points to several risk factors.
Maximum points (11) are given for active heart failure signs like S3 gallop or jugular venous distension.
Other factors include MI in the last six months, premature ventricular ectopics, rhythm other than sinus, age over 70, emergency surgery, aortic stenosis, major surgery, and poor medical condition.
A score of 0-5 predicts less than 1% risk, 6-25 predicts up to 9% risk, and over 26 predicts up to 22% risk.
What is the NSQIP risk score and how does it compare to RCRI?
The National Surgical Quality Improvement Program (NSQIP) score was developed and validated in over 200,000 surgical patients to overcome RCRI limitations.
The model includes age, ASA class, functional status, abnormal serum creatinine, and organ-based categorization of surgery.
Its predictive ability is significantly better than RCRI and works well in vascular surgical patients.
The ACS NSQIP MICA (myocardial infarction and cardiac arrest) outperformed RCRI in predicting MI and cardiac arrest combined and all-cause mortality.
What intraoperative and postoperative risk scores are available?
The P-POSSUM (Portsmouth Physiological and Operative Severity Score for the enumeration of Mortality and morbidity) has 18 variables including intraoperative characteristics like extent of surgery, blood loss, and peritoneal soiling.
Its limitation is complexity for bedside application and potential overestimation or underestimation in some populations.
The Surgical Apgar Score is a 10-point risk index considering intraoperative tachycardia, hypotension, and estimated blood loss, validated across many institutes to identify patients needing more intensive monitoring.
Risk Scoring in Cardiac Surgery
What are the main scoring systems used in contemporary cardiac surgical practice?
The two most common scoring systems are EuroSCORE 2 and STS (Society of Thoracic Surgeons) risk score.
EuroSCORE was introduced in 1999, with logistic EuroSCORE in 2006 and EuroSCORE 2 in 2011 to overcome shortcomings and keep up with contemporary practices.
The STS risk score, in addition to mortality risk, provides risk of post-operative major morbidities like renal failure, pulmonary complications, stroke, deep sternal wound infections, and re-operation probability.
What does EuroSCORE 2 measure and how was it developed?
EuroSCORE 2 predicts mortality, defined as death in the same hospital or another hospital within 30 days of operation.
It was developed studying over 22,000 cardiac surgical patients from 43 countries, with four Indian centers participating in data contribution for the first time.
NYHA classes 2,3,4 were categorized, CCS 4 was incorporated, diabetes mellitus was introduced, and creatinine clearance replaced serum creatinine as a better predictor.
The model recognizes various surgical categories like elective, urgent, emergency, or salvage surgery.
How well do EuroSCORE 2 and STS score perform in Indian patients?
Studies from AIIMS New Delhi found original EuroSCORE performed well in low and moderate risk but not in high-risk surgical populations.
Research from various Indian centers including AIIMS, Nizam's Institute, and Max Hospital found both EuroSCORE 2 and STS had satisfactory calibration power.
However, discriminatory power (differentiating between low and high-risk patients) was poor to moderate, with AUROC often less than 0.7.
A 2021 systematic review found EuroSCORE 2 calibration varied across continents, performing well in Europe and Asia but not as well in North and South America.
What are the limitations of current cardiac surgery risk scores for Indian patients?
Most Indian cardiac surgical patients (70-80%) undergo CABG, and if databases don't have adequate representation, scoring systems won't perform well.
There is very less representation of Indian cardiac surgical patients in both EuroSCORE and STS databases.
Prediction of morbidity with STS is not validated for many parameters in Indian populations.
There is a definite scope for improvement in applying these scores to Indian patients, with the most important step being forming a national database.
What factors should be considered in developing a risk score?
A clinical aim for the model must be stated first.
As many risk factors as possible should be prepared, then an appropriate modeling technique must be selected.
A suitable patient population representative of the target group is essential.
A systematic strategy to handle missing data must be adopted, and separate estimation and validation data sets are needed to get accurate numbers.
Predictive Risk Scoring in Intensive Care Units
Why are scoring systems needed in intensive care units?
Scoring systems help predict outcomes objectively, supplementing subjective physician assessment.
They assist in clinical decision-making, as stable-appearing patients with high severity scores may have poor outcomes.
They are used for benchmarking across ICUs and are considered by insurance companies and medical-legal systems.
They can predict not only mortality but also length of ICU stay, resource utilization, and duration of mechanical ventilation.
How are ICU scoring systems classified and developed?
Scoring systems are classified as first-day scoring systems (APACHE, SAPS, mortality prediction models) or repetitive scoring systems (SOFA).
Development involves collecting data from large numbers of patients across multiple ICUs, calculating scores, and most importantly, validation and customization to the population of interest.
Variables measured and timing of measurement are crucial, with most using data from the first 24 hours of ICU admission.
What is the APACHE scoring system and how has it evolved?
APACHE (Acute Physiology and Chronic Health Evaluation) was developed in 1981 with acute physiology score and chronic health evaluation.
APACHE II (1985) included 12 physiological parameters with worst value in first 24 hours, age, and chronic health points, scoring from 0-71 with mortality over 85% for scores above 34.
APACHE III (1991) studied 17,000 patients, added acid-base evaluation, and importantly considered ICU readmission and location prior to admission to prevent lead time bias.
APACHE IV (2006) studied over 130,000 patients across 104 US ICUs with 116 variables, having excellent discrimination and calibration.
What is the SOFA score and why is it important?
SOFA (Sequential Organ Failure Assessment) was developed in 1994 for use in sepsis but its applicability extended to mortality prediction.
It is a repetitive scoring system involving six organ systems, concise and useful for tracking changes over time.
If SOFA score increases after 48 hours, mortality is over 50%, even if initial score was normal.
In septic shock patients, a score over 2 has around 40% mortality compared to 10% in non-shock patients.
However, it does not take chronic health status into account.
What is the difference between discrimination and calibration in scoring systems?
Discrimination refers to how accurately a given score predicts outcome - whether a predicted mortality of 60% is actually 60%.
It is measured by area under the receiver operator curve (AUROC), with over 0.7 being good, over 0.8 very good, and over 0.9 excellent.
Calibration refers to how the scoring system performs across a wide range of predicted mortalities.
It is very sensitive to case mix and patient interventions, requiring periodic recalibration.
What are the limitations of current ICU scoring systems?
All scoring systems need external validation in the population of interest.
Some systems omit special groups like pregnancy and children.
They all need periodic recalibration, otherwise they may overpredict mortality.
Lead time bias is seen mainly with APACHE II and some SAPS models.
No single model is significantly superior to others in predicting mortality, and studies show clinical judgment combined with scoring systems predicts better than either alone.
What is the future of mortality prediction in ICUs?
Artificial intelligence and machine learning represent the future of mortality prediction.
The Superlearner Ensemble, a recent paper from the COVID era, uses 14 statistical learning models with computer-based data to predict mortality.
The MIMIC-III database, the largest ICU database in the US, has been used to develop a hybrid neural network approach called the artificial intelligence mortality score.
This system has shown superior discriminatory powers as high as 0.8-0.9 for predicting mortality at days 3, 7, and 14.
Practical Application and Summary
What risk scoring systems are recommended for different clinical settings?
For general anesthesia purposes, ASA physical status is commonly used for communication and basic categorization.
For cardiac surgical patients, EuroSCORE 2 is readily available via mobile app and provides reasonable risk estimation.
For cardiac patients undergoing non-cardiac surgery, the Revised Cardiac Risk Index (RCRI) is widely used and practical.
For critical care units, APACHE II is easily available and commonly used, while SOFA is valuable for repetitive assessment.
In resource-limited settings, APACHE II is recommended due to easy availability, while APACHE IV would be ideal with excellent calibration and discrimination.
How can clinicians incorporate functional status in risk assessment?
Functional status and capacity can be assessed through simple tests like the 6-minute walk test, which is popular and can be done anywhere.
The Duke's classification of capacity assesses how far a patient can walk and when dyspnea occurs - after 10 minutes, 4 minutes, or less than 4 minutes.
Making patients walk up to 10 meters in 4-10 minutes helps categorize them as moderate risk.
This is particularly important for cardiac patients with low ischemic threshold.
Why does ASA remain the most widely used system despite its limitations?
The ASA system "took a life of its own" and became so deeply integrated that it's difficult to remove.
Surgeons and gastroenterologists use ASA scores for monitored sedation cases.
Billing companies incorporated it into their algorithms.
It remains useful for communication between anesthesia providers and for tabular statistical grouping of patients.
However, for individual risk prediction, pivoting to other validated systems would be beneficial.
What is the role of biomarkers like BNP and troponin in risk assessment?
Pre-operative levels of BNP (brain natriuretic peptide) are independent predictors of post-operative cardiac complications.
The Canadian Cardiovascular Society guidelines recommend measuring BNP when RCRI score is greater than one or other risk factors are present.
If BNP is elevated, guidelines recommend doing troponins for 2-3 days at least once daily.
These biomarkers add important prognostic information to clinical risk scores.