The procedure for each measurement: what it is, normal values and the steps in order. Equations, pitfalls and evidence open on demand; to calculate, use the Calculator; to see cause and effect, use the Simulator. Local protocols and instructions for use take precedence.
Epicardial
Pressure wire: is the artery itself limiting flow?
a
Fractional flow reserve (FFR)
Epicardial. The fraction of normal maximal flow that is preserved despite a stenosis; a pressure ratio measured during maximal hyperemia.
Abnormal ≤0.80 · gray zone 0.75–0.80 · post-PCI <0.90 suboptimal
How to do it
- Setup: guide without side holes, transducer zeroed at heart level, line flushed and contrast cleared; administer IC nitroglycerin.
- Equalize with the sensor just beyond the guide tip (disengage the guide for ostial lesions).
- Advance the sensor to the distal part of the artery, where the vessel diameter is at least 2.0 mm by visual estimation [12].
- Induce steady hyperemia and record the lowest stable mean Pd/Pa.
- Optional pullback under sustained hyperemia; finish with a drift check at the guide tip (1.00 ± 0.02).
Equations, pitfalls and evidence
Equations
FFR = Pd,hyp / Pa,hyp
Full form: FFR = (Pd − Pv) / (Pa − Pv)
Enter your own numbers in the Calculator.
Pitfalls & artifacts
- Drift: parallel shift of Pd with intact waveforms; re-equalize if outside ±0.02.
- Damped or ventricularized Pa (contrast, deep or wedged guide): FFR falsely high. Side holes do the opposite: they make Pa look normal and mask wedging.
- Incomplete hyperemia: FFR falsely high.
- Whip, hemostatic-valve leak and transducer height errors.
- Serial lesions interact; very high microvascular resistance (e.g., acute MI) lowers maximal flow and raises FFR.
- Pressure microcatheter instead of a wire: the catheter body crosses the lesion and adds resistance, so FFR reads systematically lower (more in tight lesions); the bias disappears as it is withdrawn past each lesion. [52]
b
Non-hyperemic pressure ratios (NHPR)
Epicardial. Resting Pd/Pa, over the whole cycle or a defined diastolic or lowest-ratio window, without a hyperemic agent.
Diastolic NHPR abnormal ≤0.89 · gray zone 0.86–0.93 · resting Pd/Pa commonly ≤0.91 · post-PCI <0.95 suboptimal
How to do it
- Same setup, nitrate, equalization and sensor position as FFR.
- Wait for contrast and saline to clear: both cause transient hyperemia.
- Record several stable beats at rest; no hyperemic agent.
- Optional resting pullback to localize gradients.
- Document drift at the guide tip.
Equations, pitfalls and evidence
Equations
Resting Pd/Pa = mean Pd,rest / mean Pa,rest
Diastolic NHPR = Pd / Pa over the defined diastolic (or lowest-ratio) window at rest
Enter your own numbers in the Calculator.
Pitfalls & artifacts
- Test–retest variability is about ±0.02, so drift and noise matter most near the cutoff.
- Gray-zone values (0.86–0.93) deserve hyperemic confirmation, or contrast FFR (≤0.83).
- Arrhythmia and incomplete normalization distort diastolic windows.
When NHPR and FFR disagree (≈1 in 5 lesions)
| FFR+ / NHPR− | NHPR+ / FFR− |
|---|
| Flow | Preserved CFR; strong hyperemic response | Lower CFR; blunted hyperemia, high resting flow |
| Anatomy | Large territory (LM, proximal LAD); focal, more severe | Diffuse, less severe |
| Patient | Younger, slower heart rate | Older, diabetes, female sex, faster heart rate |
| Reading | Often behaves like concordant normal | Part of the problem may be microvascular |
- Most discordance sits near the cutoffs; check drift and hyperemia first.
- In deferred lesions, neither discordant pattern carried higher 5-year risk; only concordant abnormal did.
- Pullback (focal vs diffuse) and CFR/IMR help decide.
i
Pressure pullback: focal versus diffuse
Epicardial pattern. Withdrawing the sensor during steady hyperemia shows where along the artery pressure is lost: in one short segment (focal) or gradually along its length (diffuse).
PPG = ( Max ΔFFR20 mm ÷ ΔFFRtotal + 1 − Lfunctional ÷ Ltotal ) ÷ 2
0 = fully diffuse · 1 = fully focal · ≥0.62 predominantly focal, <0.62 predominantly diffuse · functional disease = segments losing ≥0.0015 FFR units/mm · two or more steps = serial · describe the pattern only when FFR ≤0.80
How to do it
- Sustained hyperemia (IV adenosine infusion).
- Motorized pullback at a constant speed, or a slow, steady manual pullback without stopping.
- Finish at the catheter tip and check drift.
Pitfalls, evidence and details
Pitfalls & artifacts
- IC bolus hyperemia is too short for a pullback.
- Hyperemia that wanes during pullback underestimates proximal pressure loss.
- Serial lesions interact: each step underestimates that lesion's standalone effect.
- Uneven manual speed distorts where the loss appears.
p
Post-PCI physiology: final FFR and trans-stent gradient (TSG)
Epicardial, after PCI. TSG is the drop in hyperemic FFR across the stented segment. A high TSG suggests the stent is underexpanded or undersized.
<0.01 expected in a well-expanded stent · ≥0.04 linked to more target-vessel failure and MACE · ≥0.05 action threshold used in TARGET-FFR · post-PCI FFR <0.90 suboptimal
How to do it
- Equalize, administer IC nitroglycerin, sensor distal; induce steady hyperemia and record the final FFR.
- Pull back 3–5 mm at a time and note FFR just distal and just proximal to the stent.
- Return to the guide tip for a drift check.
Practical algorithm (adapted from TARGET-FFR)
- Post-PCI FFR ≥0.90: done.
- FFR <0.90 with TSG ≥0.05: in TARGET-FFR, optimization (for example post-dilatation, guided by imaging) was considered, then FFR repeated.
- Focal step-up ≥0.05 in an unstented segment <20 mm: consider an additional stent.
- Gradual, diffuse loss without a step-up: accept the result.
Equations, pitfalls and evidence
Equations
TSG = FFRjust proximal to stent − FFRjust distal to stent · example: 0.94 − 0.88 = 0.06 (high)
Enter your own numbers in the Calculator.
Key evidence
- TARGET-FFR (randomized, n=260): 68% had post-PCI FFR <0.90 after angiography-guided PCI; physiology-guided optimization did not significantly increase final FFR ≥0.90 but reduced the rate of final FFR ≤0.80 by an absolute 11.2%.
- Registry data: TSG >0.04 predicted target-vessel failure and MACE, and in multivariable analysis TSG, not post-PCI FFR, was linked to MACE; low FFR plus high TSG carried the worst prognosis.
Caveats
- Cutoffs are not standardized (0.04 vs 0.05) and hyperemic pullbacks are noisy.
- Outcome data are mainly observational; no randomized trial has shown that lowering TSG improves clinical outcomes.
- In the lab, tick a stent in Virtual PCI (optionally "underexpanded"), re-measure and pull back to read TSG.
!
Pressure-tracing artifacts
| Artifact | How it looks | Effect on FFR | Fix |
|---|
| Signal drift | Pd shifts in parallel; waveforms and notches intact; Pd/Pa ≠ 1.00 back at the guide tip | Either direction | Re-equalize at the ostium; repeat if beyond ±0.02 |
| Damped Pa | Low-amplitude, sine-like Pa without dicrotic notch (contrast in guide) | Falsely high | Flush with saline; re-equalize |
| Ventricularized Pa | Pa loses its notch and dips in diastole, worse at hyperemia (deep or wedged guide) | Falsely high | Disengage the guide, keep the wire distal |
| Guide with side holes | Pa looks normal even when the tip is wedged: side holes mask ventricularization | Falsely high, hidden | Use a guide without side holes for physiology |
| Whip | Sharp systolic spikes on Pd at a bend | Unreliable (often low) | Move the wire a few millimetres |
| Pa leak / height error | Pa lower than cuff pressure; open hemostatic valve or transducer not at heart level | High before equalization; low if equalized on the leak | Seal the valve or remove the introducer, level the transducer, re-equalize |
| Incomplete hyperemia | No stable Pd/Pa nadir; IC bolus waning; caffeine with adenosine | Falsely high | Use IV adenosine or papaverine; confirm steady state |
| Accordion / spasm | Pseudo-stenosis from wire straightening, or vasospasm | Falsely low | IC nitrate; withdraw the stiff part of the wire |
Bolus thermodilution
3 mL saline injections: CFR, IMR and RRR
c
CFR, bolus thermodilution
Microvascular + epicardial. The ratio of hyperemic to resting flow, using mean transit time (Tmn) of room-temperature saline as an inverse flow surrogate.
Abnormal <2.5 (2.0–2.4 gray zone; older literature used <2.0) · continuous CFR runs lower: abnormal <2.0
How to do it
- Prepare: no caffeine or vasoactive drugs beforehand, coaxial 6F guide, IC nitroglycerin.
- Equalize, park the sensor at about two-thirds of the vessel, zero the temperature sensors and flush.
- Rest: three brisk 3 mL room-temperature saline injections, 20–30 s apart.
- Hyperemia: confirm steady state (mild fall in Pa and Pd/Pa), flush, then three more injections.
- Replace any Tmn more than 30% off the mean (50% if Tmn <0.25 s).
Equations, pitfalls and evidence
Equations
CFRbolus = Tmnrest / Tmnhyp
Enter your own numbers in the Calculator.
Pitfalls & artifacts
- Operator-dependent: injection speed and volume change Tmn. Console errors and fixes: slow injection (>0.6 s to nadir at the proximal thermistor) → 3 mL syringe, brisk injection with a distinct end; low amplitude (distal response <1 °C) → check guide engagement and wire position, flush with fresh room-temperature saline; no return to baseline within 8 s → wire not too distal, re-zero temperature and retry.
- Bolus values can vary by up to about 20% on repeat measurement.
- Sensor position and distal vessel volume change Tmn.
- Not specific to the microcirculation: an epicardial stenosis also lowers CFR.
- Depends on hemodynamics: high resting flow, tachycardia (about 10% lower per 15 bpm) and blood pressure.
- Higher variability than continuous thermodilution.
d
IMR, bolus thermodilution
Microvascular. Minimal achievable microvascular resistance, from the same hyperemic injections used for CFR.
IMR abnormal ≥25 · CMD (bolus) = IMR ≥25 and CFR <2.5
How to do it
- Use the hyperemic Tmn and mean Pd from the same injections as CFR.
- If FFR ≤0.80, report corrected IMR.
- Report Pd/Pa, FFR, CFR, IMR and RRR together. Low CFR + normal IMR = functional CMD; low CFR + high IMR = structural CMD.
Equations, pitfalls and evidence
Equations
IMR = Pd,hyp × Tmnhyp
IMRcorr (Yong) = Pa,hyp × Tmnhyp × (1.35 × Pd,hyp/Pa,hyp − 0.32)
IMRcorr (wedge) = Pa × Tmn × (Pd − Pw) / (Pa − Pw) (hyperemia)
RRR = (Pd × Tmn)rest / (Pd × Tmn)hyp
Enter your own numbers in the Calculator.
Pitfalls & artifacts
- Requires true maximal hyperemia; only hyperemic Tmn is used for IMR. Hyperemic agents listed in published bolus protocols: IV or IC adenosine, IC papaverine, IC nicorandil and IV regadenoson (doses per local protocol).
- IMR depends on myocardial mass and vessel volume.
- Larger distal vessel volume lengthens Tmn and raises IMR; volume-adjusted IMR agrees better with absolute resistance.
- Inherits bolus injection variability.
Continuous thermodilution
Saline infusion: absolute flow and resistance
e
Continuous thermodilution: absolute flow (Q), resistance (Rμ) and CFR
Microvascular + epicardial. Absolute volumetric flow (mL/min) at rest and at saline-induced hyperemia, measured with a constant intracoronary saline infusion.
LAD Qrest 50–100 · Qhyp 175–350 mL/min · Rμ,rest 1,200–1,800 · Rμ,hyp 200–470 WU · CFRcont <2.0 abnormal
How to do it
- Measure FFR first. Advance the infusion microcatheter (4 distal side holes) to the proximal vessel; keep the sensor distal and zero the temperature.
- Rest: saline at 10 mL/min (8 in a small vessel): record distal T, then pull the sensor into the microcatheter tip for Ti.
- Hyperemia: 20 mL/min (15 in a small vessel; saline itself induces hyperemia): record distal T, then Ti.
- Choose a format: 2-run (run 1 → stop the infusion → zero the temperature again → re-advance the sensor → run 2) or 1-run (pump programmed 10 → 20 → 10 mL/min, one pullback for both Ti values).
- Calculate Q, Rμ = Pd/Q, CFR and MRR.
Equations, pitfalls and evidence
Equations
Q = 1.08 × (Ti / T) × Qi
CFRcont = Qhyp / Qrest
Qi = saline infusion rate; Ti = infusate temperature at the microcatheter exit and T = temperature of the blood–saline mixture distally, both relative to blood temperature. 1.08 corrects for the different specific heat (and density) of saline and blood.
Rμ = Pd / Q (WU = mmHg·min/L) · Repi = (Pa − Pd) / Q
Enter your own numbers in the Calculator.
Pitfalls & artifacts
- Q describes the whole territory beyond the microcatheter tip, so place it proximal to the branches of interest.
- Stable tracings need ≥8 mL/min; 6 mL/min is unreliable.
- The microcatheter adds 2–4 mmHg of hyperemic loss, so FFR during infusion reads slightly low; measure FFR separately in the gray zone.
- Transient bradycardia or AV block, especially in the RCA; resolves on stopping the infusion.
- Values are systematically lower than bolus (mean CFR 2.67 vs 3.47), so bolus and continuous cutoffs are not interchangeable.
- Wide inter-patient ranges of Q and R: normalize to myocardial mass (for example CT-derived) for comparisons.
- Low complication rate: 0.35% of patients (0.26% of vessels) in 1,148 consecutive patients, mainly dissection, without long-term sequelae.
Reference data used to tune this lab
| Artery | Q rest (10 mL/min) | Q hyp (20 mL/min) | Rμ rest | Rμ hyp |
|---|
| LAD (n=39) | 80 ± 31 | 208 ± 74 | 1,123 ± 304 | 383 ± 140 |
| LCx (n=19) | 59 ± 16 | 145 ± 52 | 1,619 ± 467 | 663 ± 301 |
| RCA (n=30) | 66 ± 24 | 206 ± 86 | 1,467 ± 486 | 429 ± 198 |
- Flow in mL/min, resistance in WU (Gallinoro 2021). Saline at 10 mL/min left Pd/Pa (0.95 vs 0.94) and Doppler APV (22 vs 23 cm/s) unchanged; 20 mL/min lowered Pd/Pa to 0.85 and raised APV to 58 cm/s. CFRthermo 2.78 ± 0.91 matched Doppler CFVR 2.76 ± 1.06.
- Normal hearts (Fournier 2021): whole-heart hyperemic Q ≈670 mL/min, total resistance ≈150 WU; LAD hyperemic Q 293 ± 102 mL/min in controls vs 228 ± 71 with mild atherosclerosis (difference driven by epicardial resistance). Per gram, hyperemic flow is similar across territories (≈5–6 mL/min/g).
- ANOCA cohorts: CFRcont 2.6–2.7 and MRRcont ≈3.2 (bolus 3.3–3.5 and ≈4.4). Test–retest variability ≈12–13% for continuous CFR and MRR vs ≈31% for bolus CFR and ≈24% for IMR.
- Patterns: functional CMD = high resting Q, low resting Rμ (for example HFrEF); structural CMD = high hyperemic Rμ, low hyperemic Q (for example HFpEF). Diabetes lowers CFR (2.38 vs 2.88) and MRR (2.79 vs 3.48) with similar absolute flows.
- Validation: swine flow probes (r = 0.84 for flow), [15O]H2O PET (r = 0.91), saline vs adenosine hyperemia (r = 0.97); a programmed 1-run 10 → 20 → 10 mL/min sequence agrees with the 2-run method.
- In this lab, LCx and RCA territories are scaled from these per-vessel means. Bolus and continuous share one simulated physiology, so the lab does not reproduce the systematic bolus-versus-continuous offset seen clinically.
MRR
The microcirculation on its own, from either thermodilution method
M
MRR: microvascular resistance reserve (bolus or continuous)
Microvascular-specific. How much the microcirculation alone can dilate, with the epicardial stenosis and the blood-pressure change removed. Full equation recommended; see the Calculator for a worked example.
≥3.0 makes CMD unlikely · 2.3–3.0 gray zone · <2.3 CMD likely
How to do it
- Run the bolus or continuous protocol; record Pa at rest, and Pa and Pd at hyperemia.
- Take FFR from the same hyperemic run.
- Calculate MRR with the full equation.
Equations, pitfalls and evidence
Equations
MRR = (CFR / FFR) × (Pa,rest / Pa,hyp) = (Qhyp / Qrest) × (Pa,rest / Pd,hyp)
Simplified: CFR / FFR, valid only when Pa does not fall (saline hyperemia)
Enter your own numbers in the Calculator.
- Theoretically independent of epicardial stenosis and myocardial mass, so more specific than CFR or IMR.
- Bolus MRR runs higher than continuous (mean 4.40 vs 3.22) with about 3-fold worse repeatability; the 2.3 and 3.0 thresholds come from continuous thermodilution, so apply them to bolus values cautiously.
- For continuous, use Pa at 10 mL/min (rest) and at 20 mL/min (hyperemia).
- Thresholds come from exploratory analyses; outcome data are accumulating.
E
Microvascular endotypes · what the numbers mean
| Endotype | CFR | IMR / HMR | Acetylcholine | What is going on |
|---|
| Functional CMD | <2.5 | normal (IMR <25 · HMR <2.5) | — | Resting flow is high (the small vessels already work hard at rest); they still dilate normally, so the reserve is used up. |
| Structural CMD | <2.5 | high (IMR ≥25 · HMR ≥2.5) | — | The small vessels cannot open fully: minimal resistance is high and peak flow is low. |
| Endothelial dysfunction | any | any | flow rises ≤1.5× | The vessel lining fails to release nitric oxide; the flow response to acetylcholine is blunted. |
| Microvascular spasm | any | any | angina + ischemic ECG change, no epicardial spasm (flow often falls: supporting, not required) | The small vessels clamp down. |
| Epicardial spasm | any | any | ≥90% constriction + angina + ECG change | The large artery clamps down. |
| Non-cardiac | ≥2.5 | normal | normal | All tests normal: look elsewhere. |
The Simulator shows functional and structural CMD (vasodilator physiology); acetylcholine endotypes are tested in the Doppler module. Thresholds: CFR <2.5 bolus/Doppler (continuous CFR runs lower, abnormal <2.0). [62] Rahman 2019 · [63] Rahman 2020 · [64] Perera 2023
Doppler
Flow velocity wire: CFR, HMR, HSR and acetylcholine testing
j
Doppler: CFR, HMR, HSR and acetylcholine testing
Endothelium-independent (adenosine) and endothelium-dependent (acetylcholine) coronary function from a Doppler (± pressure) sensor-tipped wire, measured in the proximal-mid vessel.
CFRne >2.5 normal (often 3–5 in normal arteries; older studies <2.0 abnormal) · HMR normal <2.0–2.5 · HSR >0.8 ischemic · ACh: CBF rise ≥50% normal · epicardial constriction >20% abnormal · spasm ≥90% with symptoms and ischemic ECG · NTG dilation >20% normal
How to do it
- Prepare: withhold vasoactive drugs and caffeine, heparinize, LAD by default; no nitroglycerin before testing.
- Infusion catheter in the proximal–mid vessel, Doppler wire just distal and free in the lumen; optimize the envelope and record resting APV.
- IC adenosine, repeated until peak APV is stable: CFR, HMR and HSR.
- Acetylcholine low → intermediate → high (about 3 min each): APV and QCA diameter, ECG and symptoms; stop for symptoms.
- IC nitroglycerin at the end (or at once for spasm).
Equations, pitfalls and evidence
Equations
CFRDoppler = APVhyp / APVrest
HMR = Pd,hyp / APVhyp (mmHg·s/cm) HSR = (Pa − Pd)hyp / APVhyp
CBF = 0.5 × APV × π r² ΔCBFACh = CBFACh / CBFrest − 1
Enter your own numbers in the Calculator.
Pitfalls & artifacts
- Poor or variable signal underestimates APV; never record before the envelope is stable.
- Doppler velocity changes reflect flow only if the vessel area at the sample site is unchanged; during ACh, compute CBF from APV and the QCA diameter (a constricted segment raises velocity at unchanged flow).
- Widmer et al. describe a >20% diameter increase and/or >50% CBF increase with ACh as normal; most protocols grade epicardial constriction >20% as endothelial dysfunction.
- Operator-dependent, with the highest variability of the wire methods.
- Acetylcholine can cause bradycardia/AV block (especially RCA) and spasm; have nitroglycerin ready. Major complications are uncommon (about 0.5–1%).
- Microvascular spasm = reproduction of angina with ischemic ECG changes but <90% epicardial constriction.
Other indices
rCFR = CFRtarget / CFRreference (≈1.0) · FFRcor = (Pd − Pw) / (Pa − Pw) · AchFR = QAch / Qrest (<1.5, exploratory)
∑
Teaching model behind the simulator
Stenosis: ΔP = f·Q + s·Q² f = 8πμL / As² s = (ρ/2)·k·(1/As − 1/An)²
Circuit: Pa − Pv = Q·(Rμ + Σf) + Σs·Q² Pd(x) = Pa − losses proximal to x
Thermodilution: Tmn = V / Q (V ≈ 0.6 mL) IMR = Pd × Tmn
Gould-type losses scaled to clinical FFR values, autoregulated resting flow, minimal microvascular resistance at hyperemia. Serial lesions share one flow, which is why each masks the other. Teaching approximations only.
