How to Read an ACC/AHA Guideline Recommendation
Lesson 4 Supplement — Cardiovascular Pharmacology
How to Read an ACC/AHA Guideline Recommendation
Every cardiology guideline from the ACC and AHA rates each recommendation on two independent dimensions: how strongly they recommend it (Class of Recommendation, COR) and how good the evidence is (Level of Evidence, LOE). Both appear every time — together, they tell you how much weight to give a recommendation in clinical practice.
Part 1 — Class of Recommendation (COR): “How Strongly?”
The COR tells you whether you should do something, consider doing it, or avoid it altogether.
| Class | Plain Language | What It Means Clinically |
|---|---|---|
| Class I | Should do this | Strong evidence or agreement that benefit outweighs risk. Do it unless contraindicated. |
| Class IIa | Reasonable to do | Weight of evidence favors benefit. Most patients will benefit — sound choice in the right patient. |
| Class IIb | May consider this | Benefit less certain. Evidence is weaker or conflicting. Use clinical judgment. |
| Class III: No Benefit | Don’t bother | Evidence shows no benefit in this specific indication. Doing it is not harmful — but it doesn’t help. |
| Class III: Harm | Don’t do this | Evidence of harm. Doing it in this situation is actively dangerous. |
The language cue: Guideline wording is not accidental. “Is recommended” = Class I. “Is reasonable” = Class IIa. “May be considered” = Class IIb. “Should not be done” = Class III. Once you know the translation, you can read the class without looking it up.
Part 2 — Level of Evidence (LOE): “How Good Is the Data?”
The LOE tells you what kind of research supports the recommendation. A weak LOE does not automatically weaken the recommendation — it may simply mean that a large RCT has never been done or is not feasible.
| Level | What It Means |
|---|---|
| A | Multiple high-quality RCTs, or meta-analyses of high-quality RCTs. The gold standard. |
| B-R | At least one moderate-quality RCT, or meta-analyses of moderate-quality RCTs. Solid randomized data, more limited scope. |
| B-NR | Well-designed nonrandomized studies, observational data, or registry studies. Good data, but not randomized. |
| C-LD | Randomized or nonrandomized studies with significant design limitations, or mechanistic/physiological studies. |
| C-EO | Expert opinion based on clinical experience. No direct trial data exists — this is the panel’s best collective judgment. |
Context matters: Many Class I recommendations carry Level C-EO — particularly for procedures where a placebo-controlled trial would be unethical. That does not make them optional. “Shock a patient in pulseless VF” is Class I, Level C-EO because nobody ran a randomized trial.
Part 3 — Reading a Recommendation
- Class I — You should do this. Strong recommendation.
- Level A — Multiple high-quality RCTs support it (JUPITER, 4S, HPS, ASCOT-LLA, CARDS, etc.).
- “Unless contraindicated” — The only acceptable reason not to follow Class I is a specific contraindication.
- Class IIb — May be considered. Weaker recommendation — benefit is not clearly established.
- Level B-R — One moderate-quality RCT (TOPCAT) with limitations. Real data, but not definitive.
Class I + Level A is the gold standard of guideline-based care. If a patient with a Class I, Level A indication is not on that therapy, you need a documented reason why.
This is one of 13 free reference sheets from the APP Cardiology Academy — no account required.