It teaches how contraction shapes coronary flow across the wall and what diastolic time does to it, in an adult and animal model with no paediatric parameter. It cannot show ischaemia, oxygen demand or the layers autoregulating separately.
A thick-walled left ventricle in a closed circulation, with a coronary bed split into three layers at relative depths 1/6, 3/6 and 5/6 of the wall. It is integrated here in your browser and redrawn every frame; the sliders take effect in the middle of a beat.
Each layer's intramyocardial pressure is its depth multiplied by cavity pressure, plus a varying-elastance term. That pressure squeezes the layer's microvascular compartment, and a smaller compartment has a higher resistance (resistance rises as the inverse square of volume), so flow into the deepest layer falls and can reverse in systole.
Aortic diastolic pressure and LVEDP are targets, not inputs: the circulation is closed, so they are reached gradually by changing systemic resistance and preload. Both the target and the value actually achieved are shown.
Fitted to 1.11 at rest.
The invasive quantity.
Fitted to 2.2 at rest.
Fitted to 1.11 at rest, so whether it is above 1 comes from the fit.
Read the direction, not the level or the size: both depend on assumptions listed below.
Computed from modelled aortic and LV pressures, so this is the invasive quantity. No threshold is printed: the sources disagree.
This model rests near 1.9, above every invasive value we read, so read the direction, not the level.
Fitted to 2.2 at rest.
Subendocardial / subepicardial flow.
Supply : demand, DPTI/SPTI.
| Layer | Depth | Mean flow mL/s | Mean flow mL/s/100 g | Minimum mL/s | Retrograde ms | Peak pim mmHg | Arteriolar calibre |
|---|---|---|---|---|---|---|---|
| Measuring… the first beat has not closed yet. | |||||||
Experimental teaching model. The base model reproduces its published reference beat, but no parameter in it is paediatric human, the ventricle and circulation driving it are not validated above about 115 bpm, and there are no published transmural coronary flow measurements in children at all. It is not a patient model, it makes no claim to paediatric quantitative accuracy, and it carries no treatment recommendation. Read the direction, not the size: the size depends on assumptions listed below. The direction comes from one prescribed rule, not real anatomy: a layer's intramyocardial pressure is its relative depth times cavity pressure. The resting ratio is fitted, and the fit reverses it. Without three frozen per-layer resistance factors this model rests at 0.70, with the subendocardium below the subepicardium; the factors lift 0.70 to 1.11. The fitted factors set the level, so whether the ratio is above or below 1 under load comes from the fit and is not a result: at 140 bpm it reads 0.94 with them and 0.59 without.
A thick-walled spherical left ventricle with fibre and radial wall stress drives a layered coronary circuit inside a closed windkessel circulation. Thirteen ordinary differential equations are integrated in your browser with a fourth-order Runge–Kutta step of 0.2 ms; the page draws every frame, so the sliders act in the middle of a beat. The state is the cavity volume, ten compartment volumes (systemic arterial and venous, epicardial coronary arterial and venous, and an arteriolar and a venular compartment in each of the three layers) and two valve flows. The unlayered base model, which is what the published reference beat is checked against, has eight: the same without the three layers.
The base model is Bovendeerd 2006, taken as printed. It was chosen because it is the only one of the four candidate models that can be reproduced from its own paper: all 22 equations and both parameter tables are published with units. It has no transmural layers, and says so: the flow in it “should be regarded as a mean flow over the myocardial wall”.
The three layers are Munneke 2022's, grafted on. That model has the layers and the only validation against an independent clinical condition (coronary velocity before and after aortic valve replacement), but it rides on a host circulation model whose parameter values are not in the paper, so it is cited as the authority for the structure and for the validation numbers and is not used as the code base. Mynard 2014, which supplies the layer splits and gradings, has the stronger animal validation of the two – in vivo in nine adult sheep and nine newborn lambs – but it is not closed-loop: measured aortic and ventricular pressures are its inputs, so it cannot be driven by sliders, which is the same objection that rules out the autoregulation model as a base. Bovendeerd's own single-radius intramyocardial pressure equation is Munneke's transmitted-cavity-pressure term evaluated at one radius, so replacing one radius with three is the same equation evaluated three times.
| Quantity | Relation | Source |
|---|---|---|
| Cavity pressure | plv = (σf − 2σm,r) · ln(ro/ri) | Bovendeerd Eq 7 |
| Fibre stretch, sarcomere length | λf = [(Vlv + Vw/3)/(Vlv,0 + Vw/3)]1/3 | Bovendeerd Eq 8 |
| Active stress | σa = c · σar · f(ls) · g(ta) · h(vs), with g = sin²(πta/tmax) | Bovendeerd Eq 10–13 |
| Passive fibre and radial stress | σm = σ0(ec(λ−1) − 1) for λ ≥ 1, else 0 | Bovendeerd Eq 14–15 |
| Intramyocardial pressure, one radius | p̄im = σ̄m,r + [(ro − r̄)/(ro − ri)] · plv | Bovendeerd Eq 16–18 |
| Intramyocardial pressure by layer | pim,j = σ̄m,r + rj · plv + γσf, with rj = 1/6, 3/6, 5/6 and γ = 0.06 | Munneke Eq 1; the same depths in Gharahi Eq 3 and Mynard |
| Compartment volume | Vi = max(V0,i + Ci · ptm,i, Vc) | Linear law: Gharahi Eq 2, Mynard Eq 5. The floor: Gharahi only |
| Compartment resistance | Ri = R0,i · V0,i²/Vi² | Munneke Eq 3, Gharahi Eq 1, Mynard Eq 6 |
| Shared middle resistance | Rm = R0,m[γV0,1²/V1² + (1−γ)V0,2²/V2²], γ = 0.75 | Mynard Eq 7 |
| Compliance, resistance, inertance | Δp = (V − V0)/C; Δp = Rq; Δp = L dq/dt | Bovendeerd Eq 19–21 |
| Epicardial coronary stenosis | Rs = R0α−2, α = area ratio | Duanmu Eq 1 |
| Distal / proximal pressure ratio | pd/pa | Duanmu Eq 2 |
| Supply : demand (the subendocardial viability ratio, SEVR, or Buckberg index) | DPTI/SPTI, DPTI = ∫(pao − plv)dt over diastole, SPTI = ∫plvdt over systole | DPTI: Buckberg 1972. SPTI from LV pressure: Hoffman & Buckberg 2014 |
Ventricle, from Bovendeerd Table 1: wall volume 200 mL, zero-pressure cavity volume 60 mL, sarcomere length 1.9 µm at unit stretch, passive fibre stress scale 0.9 kPa with exponent 12, passive radial stress scale 0.2 kPa with exponent 9, active stress scale 55 kPa, activation duration 400 ms, unloaded shortening velocity 10 µm/s, Hill curvature zero. Bovendeerd takes the 200 mL wall as 200 g, which is the mass behind every per-100 g figure here.
Circulation, from Bovendeerd Table 2, in 106 Pa·s·m−3 and 10−9 m³·Pa−1: systemic resistances 5, 120, 5; compliances 20 and 800; unstressed volumes 500 and 3000 mL; inertances 60×10³ Pa·s²·m−3; blood volume 5000 mL. Coronary resistances 700, 900, 900, 200 normally and 200, 100, 100, 200 at maximal vasodilation; compliances 0.03, 1.4, 0.7; unstressed volumes 6, 7, 10 mL. Those three compliances are Spaan 1985's 0.0022, 0.091 and 0.045 mL·mmHg−1·100 g−1 for large arteries, the myocardial bed and large veins. They are not measurements: Spaan's table is headed as an estimate based on assumptions explained in his text, it is built mostly from dog data, and he footnotes the venous value as one to apply with care because it rests on assuming venous volume is twice arterial volume. His own conclusion is that little adequate information exists on blood volume distribution in the heart and still less on the wall properties of the myocardial microcirculation. The total coronary resistance matches the 66 mmHg·min·g·mL−1 and the 25 : 68 : 7 % split that Bovendeerd quotes from Chilian 1989, in anaesthetised cats abstract.
Layers: resistance split within the intramyocardial compartment 60 : 30 : 10, quoted by Munneke from Chilian 1989 secondary, though it is not in the abstract we read; arteriolar to venular split of compliance about 5 : 95 and of unstressed volume about 24 : 76, and a 1.14 subendocardial-to-subepicardial grading of all four, all taken from Mynard 2014 secondary. Mynard measures none of them. It cites the splits to Spaan 1985 and to Spaan 2000 (canine, not read), and the 1.14 to a 1988 mathematical model abstract and a 1974 paper with no abstract, in dogs by PubMed's subject headings. The grading does almost nothing here: applying one factor to both compliance and unstressed volume cancels in the resistance law, so it moves the resting ratio about 0.3% and the subendocardial retrograde minimum about 3%. The collapse floor is 0.43 of the unstressed volume, the ratio of two values fitted to a porcine LAD bed. Gharahi states that floor on volume. That a compartment at it then passes flow in series is our deduction, by conservation of mass.
The direction comes from one prescribed rule, not real anatomy: a layer's intramyocardial pressure is its relative depth times cavity pressure. Flatten that rule and the ratio does not fall with heart rate at all. How far it falls with heart rate is set mostly by the compliance: at 150 bpm ×0.83 here, ×0.78 with Mynard's full set, ×0.64 with the compliance alone raised 2.94-fold. The floor matters more for aortic stenosis and a lesion. The direction held across every closure tried for heart rate, LVEDP, aortic stenosis and a lesion; the size did not.
Without three frozen per-layer resistance factors this model rests at 0.70, with the subendocardium below the subepicardium. Hoffman and Buckberg give about 1 to 1.2 as normal. The factors span 1.57-fold and lift 0.70 to 1.11, a target that traces to a textbook. They sit where resting transmural tone would, which is our reading and not a sourced statement, and they do not behave like tone: they stay 1.57-fold at maximal vasodilation, so at rest subendocardial reserve is no lower than subepicardial here.
Two things are fitted once, at the reference beat, and then frozen as constants in the code: one global factor on the intramyocardial resistances, so that resting coronary inflow returns to Bovendeerd's own 2.2 mL·s−1·100 g−1, and three per-layer factors (1.255, 0.996, 0.800, outer to inner), so that the resting subendocardial/subepicardial flow ratio equals 1.11. The mid-wall share of resting flow is the geometric mean of the other two, which is an assumption of the fit, so every resting layer flow in the table is set by it. Every scenario then runs with those same numbers.
The table separates what the fit does from what the depth rule does. Each cell is the ratio, and in brackets its multiple of that row's resting value, with the page's diastolic pressure and LVEDP targets held.
| Configuration | Rest | 140 bpm | 160 bpm | LVEDP 20 | Aortic stenosis |
|---|---|---|---|---|---|
| As shipped | 1.11 | 0.94 (×0.85) | 0.90 (×0.81) | 1.03 (×0.93) | 1.04 (×0.94) |
| Fitted factors out | 0.70 | 0.59 (×0.84) | 0.56 (×0.80) | 0.65 (×0.93) | 0.66 (×0.94) |
| Depth rule flat | 1.56 | 1.56 (×1.00) | 1.56 (×1.00) | 1.56 (×1.00) | 1.59 (×1.02) |
The authors fit too. Munneke distributes its target flow with a ratio of 1.11, from a textbook (Crawford 2009), and adjusts a reference pressure drop to reach it, calling the result an auto-regulated healthy control simulation. Mynard assumes ratios measured in sheep and determines total resistance iteratively. Gharahi penalises departure from 1.25, which it calls the approximate average of values reported in the literature, and prints its factor: subepicardial-to-subendocardial resistance 1.45 (range 1.006 to 2.05). We did not find either of the first two printing what its ratio would be without the fit, so we cannot say theirs does less. What differs here is only that we measured the untuned value and said so.
At rest these factors occupy the place of transmural tone. That is our reading, not a sourced statement, and they do not behave like tone: they stay 1.57-fold at maximal vasodilation, and the ratio rises as the bed dilates, from 1.11 at normal tone to 1.15 fully dilated. So at rest this subendocardium has no less reserve than its subepicardium. Away from rest, even with no lesion, that can reverse: a low diastolic pressure with a fast rate or a high LVEDP. Letting the factors fade with tone would be a change to the model, which we have not made.
A compartment at the floor conducts at a fixed 5.5-fold resistance however hard it is squeezed, so further compression stops costing anything. Munneke's law has no hard floor: as volume approaches collapse, transmural pressure becomes progressively negative, and resistance goes on rising. Mynard has no floor either. Only Gharahi has one, fitted to porcine LAD beds. At rest here the subendocardial venular compartment is at the floor for 13% of each beat, and its pressure reaches −14 mmHg.
Each row below is refitted to 1.11 and 2.2, with the page's diastolic pressure and LVEDP targets held, so each multiple is of that row's own rest. The direction of the first four columns holds in every row and the size does not. The direction of the last column does not hold either.
| Closure | 150 bpm | LVEDP 20 | Aortic stenosis, c 1.25 | 85% lesion | Aortic stenosis ratio | Steady dilated inlet 100 / 60 / 45 mmHg |
|---|---|---|---|---|---|---|
| As shipped, floor 0.43 of V0 | ×0.83 | ×0.93 | ×0.94 | ×0.82 | 1.04 | 1.22 / 1.25 / 1.29 |
| Floor 0.60 | ×0.85 | ×0.93 | ×0.96 | ×0.82 | 1.06 | 1.19 / 1.22 / 1.25 |
| Floor 0.30 | ×0.84 | ×0.93 | ×0.92 | ×0.77 | 1.02 | 1.23 / 1.26 / 1.31 |
| Floor 0.20 | ×0.84 | ×0.93 | ×0.81 | ×0.69 | 0.89 | 1.19 / 1.22 / 1.27 |
| Floor 0.05 | ×0.84 | ×0.93 | ×0.78 | ×0.69 | 0.87 | 1.17 / 0.99 / 0.97 |
| Floor 0.43, Mynard's compliance, unstressed volume and resistance split | ×0.78 | ×0.92 | ×0.77 | ×0.65 | 0.86 | 1.28 / 1.28 / 1.33 |
| Floor 0.43, compliance alone ×2.94, V0 unchanged | ×0.64 | ×0.85 | ×0.68 | ×0.49 | 0.75 | 1.41 / 1.44 / 1.50 |
The last row is a scan and not Mynard's parameter set, which has its own unstressed volumes. 2.94 is about the ratio of Mynard's total compliance to the one used here, 0.267 against 0.091 mL·mmHg−1·100 g−1.
Invasive values in children mostly sit around 0.9 to 1.0 (the cohorts are under Sources), and tonometric values around 1.1 to 1.5. The published cut-offs, 45% invasive and 130% non-invasive, belong to those different scales. The human series collected in Hoffman and Buckberg's Table 1 secondary are nearly all patients with aortic valve disease, about 0.5 to 0.9 without ischaemia, uncorrected for oxygen content. The widest range, 0.43 to 1.68 in their table and 0.43 to 1.65 in the paper's abstract abstract, is asymptomatic children with congenital aortic stenosis. Children with normal aortic valves in that paper averaged 1.27 (0.96 to 1.78), with a denominator taken over ejection only. We read no invasive resting value in normal adults.
This model's 1.9 is above all of these. Part of that is its short systole: 34% of the beat, against 41% reported for the other adult model, and at 41% it would be about 1.4. The rest we have not explained.
The demand side of the index, SPTI, is a pressure integral and nothing else. Hoffman and Buckberg write of it, as an estimate of oxygen use that comes from Sarnoff's time-tension index: “This index is not ideal because myocardial oxygen usage is more closely related to wall stress than to pressure and because it does not take muscle mass or contractility into account.” On the supply side, the same authors write that DPTI, even corrected for oxygen content, is “still an index and not a true oxygen supply, because it does not consider the vascular resistance”. In anaesthetised dogs, isoproterenol raised the critical ratio from 0.8 to 1–1.1, as they report. So a rise on the tile after an inotrope is not better balance: the tile says so whenever contractility is off its reference or the valve is narrowed, and gives no direction then. It gives none with a coronary lesion either, because the index cannot see one.
The diastolic pressure slider lowers systemic resistance, because the circulation is closed. Taking the target from 82 to 45 mmHg takes peak LV pressure from 118 to 91 mmHg. Flow falls 44%, 43% and 43% from outer to inner layer, and the ratio rises 2.5% (3.6% with the fitted factors out, so the fit is not the reason). That is vasodilated hypotension. The run-off of a duct, a shunt or aortic regurgitation keeps LV pressure up, which this model does not cover.
The rise is the LV term winning. With the bed run alone on the page's own recorded waveforms, the low inlet with the resting ventricle lowers the ratio 4.9%, the resting inlet with the low ventricle raises it 9.3%, and the two together raise it 2.6%. A steady inlet is not a fair version of this: taken from 97 to 37 mmHg it moves the ratio +0.5%.
What has been measured, and what this model does in the same experiment:
| Experiment | Source | This model | Scored |
|---|---|---|---|
| Main stem perfused at controlled pressure, full adenosine dilation | Goats: 1.56, 1.39, 1.22, 1.07 at 102, 82, 62, 47 mmHg, diastolic time fraction 0.66. Our arithmetic from the two printed regressions: the abstract reports Cepi itself as not significant, so its slope here is recovered from the DTF1 regression; with Cepi flat the direction is the same (Fokkema 2005 abstract) | Steady inlet: 1.22, 1.23, 1.25, 1.29 at 100, 80, 60, 45 mmHg | Wrong direction: fails |
| Same, diastolic time fraction lowered | Goats: subendocardial conductance falls with it (Fokkema 2005 abstract) | 1.22 to 1.09 from 0.66 to 0.50 | Right direction: passes |
| Occluder under adenosine, distal pressure about 70, 50, 35% of control | Dogs: the ratio falls (r = 0.83); zero-flow pressure 10 mmHg subepicardially, 18 subendocardially (Bache & Schwartz 1982 abstract) | 1.15, 1.08, 1.04, 0.98 at 90, 63, 46, 31 mmHg | Right direction: passes |
| Spontaneous autoregulation, pressure lowered | Conscious dogs: endocardial flow flat from 84 to 49 mmHg, lower limit about 40 (Canty 1988 abstract); failure below about 38 mmHg subendocardially and 28 subepicardially (Hoffman & Buckberg 2014, Fig 3) | Autoregulation on, diastolic 25 mmHg: ratio 1.15, no fall. One gain, and factors that do not fade, so this does not appear at this setting | Not scored: out of scope |
| Run-off: an arteriovenous fistula opened progressively | Anaesthetised open-chest dogs: aortic diastolic pressure fell while systolic pressure held until near the largest flows, and the ratio fell with diastolic flow (Buckberg 1972) | Not covered: the slider lowers systemic resistance, with no run-off | Not scored: out of scope |
| A 60 mmHg reduction in perfusion pressure, in a network simulation that opens with stenosis and attributes its fall mainly to non-linear compliance | A 20% fall in the ratio (Algranati 2011 abstract). A simulation, not a measurement | Lesion taking about 60 mmHg off the bed: −18.5%. The page's low inlet alone, about 34 mmHg off the mean: −4.9% | Lesion passes; low inlet fails |
So a lesion moves the ratio the way it was measured, and a lowered inlet under dilation does not. With the floor lowered to 0.05 the steady dilated inlet does fall (the table above).
Smaller assumptions: the three layers take equal thirds of the compartment by volume while their depths are thirds by thickness (in a sphere those are not the same split, and no source states a mass split); the coronary lesion is a 10 mm segment of a 4.24 mm vessel, since no source gives a lesion length; the ceiling on vasoconstriction above resting tone is ours, because no source quantifies it; the autoregulator's update has a floor of 0.02 on tone, also ours, which among reported results changes only the 78% lesion figures, by under 2 percentage points; and the radial wall-stress term in intramyocardial pressure is the base model's single mean value, so it is the same in all three layers rather than depth-resolved (it is zero whenever the cavity sits above the passive zero-pressure volume, so it only matters for empty or very low-afterload beats).
The valve-area slider covers the case Munneke validated, 4.8 to 0.8 cm², and goes on to 0.5 cm² with a beyond-range caution. Munneke's valve law is not printed in the paper or its supplement, and Bovendeerd's valve is a bare diode with an inertance, so the gradient here is the convective (simplified Bernoulli) term ρq²/2A², written as the difference from the 4.8 cm² reference so that a normal valve adds nothing. It is not a Gorlin effective orifice: there is no discharge coefficient and no pressure recovery.
The coronary stenosis slider is labelled in diameter, with the area ratio shown beside it, because the source for the scaling law warns that clinical stenosis severity is a radius ratio and its own law is in area. Bovendeerd has one coronary bed for the whole wall, so a lesion in series with it behaves like a left main lesion feeding everything. Pd/Pa is the pressure just distal to the lesion over the inlet pressure, the paper's definition, so it reads 1.00 with no lesion. It is taken at the flow of the beat shown: resting tone with autoregulation off, near maximal dilation behind a severe lesion with it on. Either way it is not a fractional flow reserve: this lesion is a Poiseuille resistance with no expansion loss, feeding the whole wall. Near no flow the ratio means little: at 85% diameter stenosis, autoregulation off, 19% of forward arteriolar flow goes back out to refill another layer, and at 90%, 60%. The page says so when that share passes a quarter.
Recorded under verification/ in this site's repository and rerunnable. Scored checks are reported by kind. Base model against its own paper: 28 of 28. Layered model against published models and measurements: 20 of 37. A further 11 rows cannot fail on their own (a calibration target, a constant, a symmetry, the paper's own sum) and 5 are forced by how the code is written (our activation law, the autoregulation gain, Poiseuille with an assumed lesion length), so neither set is counted. The other 67 rows are this model's behaviour on the record, quantities outside its scope, and comparisons that are not like for like. Some of those last would fail if scored.
Reproduced. The reference beat, within 5%: peak LV pressure 15.8 against 16.1 kPa, peak intramyocardial pressure 8.90 against 9.1 kPa, mean coronary inflow 2.13 against 2.2 mL·s−1·100 g−1, coronary perfusion pressure 11.0–16.0 against 11.4–16.3 kPa, myocardial coronary pressure 2.9–12.0 against 3.0–12.3 kPa, myocardial coronary volume 10.6–12.3 against 10.7–12.5 mL, stroke volume 77 against 75.6 mL, ejection fraction 65 against 64%, peak aortic and mitral flows 654 and 313 against 646 and 307 mL/s, left atrial pressure 1.42 against 1.45 kPa. The two-peaked inflow with diastolic predominance and the single systolic outflow peak are reproduced.
In the constant-10-kPa maximal-vasodilation condition, peak inflow is 14.2 against 14.2, bed volume 5.8–12.5 against 5.8–12.4 mL, and the early-systolic inflow peak disappears as reported. The mean, 8.33 against 8.3, is not a test: it is the paper's own sum, 10 kPa over 6×108 Pa·s·m−3 and a 200 g wall, and the base model's resistances are constant. The slight retrograde minimum is not reproduced: −0.1 published, +0.02 here.
The layered model's own waveform-shape metrics are per named epicardial vessel and in velocity, and this model has one coronary bed and no vessel geometry, so they are reported rather than scored: the peak diastolic-to-systolic inflow ratio is 2.3 against a published 1.2 to 2.3 by branch, and the diastolic share of total inflow is 0.88 against a published 0.7 to 0.8.
Of the base model's seven named experimental observations, five are reproduced (flow amplitude vanishing at zero contractility and rising with it; amplitude nearly flat below 13 kPa of cavity pressure; minimum systolic flow nearly flat below it; minimum systolic flow falling above it; endocardial intramyocardial pressure exceeding cavity pressure at low afterload). One needs a clamped preparation this closed circulation does not have, and one the original paper could not reproduce either. The first is reproduced only in direction: amplitude rises with contractility sub-proportionally (4.15, 3.70 and 2.92 per unit contractility at 0.25, 0.5 and 1), not in proportion as reported.
The layered behaviour matches the reported transmural pattern: subendocardial flow is the most pulsatile and goes retrograde in systole, and subepicardial flow does not reverse and has similar magnitude in systole and diastole (mean 1.33 against 1.42 mL/s, peak 1.75 against 1.58). Mid-wall flow is mostly diastolic, 81% by volume, which is looser than the reported “almost all”. Tachycardia lowers the ratio. Raising LVEDP from 10 to 20 mmHg with mean pressure held cuts supply:demand 17% against a reported 20%; with mean pressure free it is 9%, and the source line we have says nothing about holding it.
Aortic stenosis: a prediction, a validation and a measurement. Munneke's fall in the ratio from 1.11 to 0.38 and 0.39 in the two left branches is a model prediction for a non-chronic lesion with no adaptation, not a measurement. In anaesthetised open-chest dogs, tightening a supravalvular aortic constriction, among other procedures, lowered the subendocardial:subepicardial microsphere flow ratio, and ratios below 0.8 always occurred when DPTI/TTI was below 0.8 (Buckberg 1972). What Munneke validated against adult patients is LAD velocity before and after valve replacement. Transmural flow has also been measured in humans: by PET, in 15 adults with the most severe stenosis, the ratio was 1.14 at rest and 0.92 at hyperaemia, and reserve was 1.43 subendocardially against 1.78 subepicardially (Rajappan 2002 abstract). Those adults have hypertrophy and this case has none, so only direction is scored.
Against the measurement. At rest the model's 1.04 to 1.07 sits near the measured 1.14, but that level comes from the fit. The real gap is dilation. The ratio rises here, 1.04 to 1.11, where it fell in adults, and subendocardial reserve is above subepicardial here (5.06 against 4.73) where it was below. Both rows fail.
Against Munneke. Every row is scored at both contractilities, so no row is scored only where it passes, with mean pressure held as Munneke holds it. Peak systolic inflow falls to ×0.79 and ×0.84 of rest against about ×0.31 (LAD velocity 0.13 to 0.04 m/s in the model, 0.16 to 0.05 measured in adults, as Munneke reports): fails. Peak diastolic inflow is ×1.00 and ×1.04, and it was unaffected in both Munneke's model and the measurement: passes. Mean flow falls 7% and 8% against 35%, and the ratio stays at 1.07 and 1.04 against 0.38: both fail. Munneke also finds reverse systolic flow; here the lowest inflow in ejection is 0.49 mL/s and −0.03 mL/s.
A likely reason, not run here: Munneke's compartments follow a printed collapsible tube law whose intramyocardial reference pressure is not printed, so it could be tested by sweeping that pressure. This version does not. With mean pressure held, raising the compliance alone reaches 0.75 at 2.94 times, the same as the floor table with diastolic pressure held, and 0.49 at six times the value used here, which no source supports.
The valve itself. This ventricle cannot match peak cavity pressure, mean gradient and cardiac output at once through a 0.8 cm² orifice. At the base model's own contractility peak pressure passes (159 against 160 mmHg), but the mean gradient (32 against 45 mmHg) and output (3.3 against 5.0 L/min) fail. Raising contractility a quarter passes the gradient (47 mmHg) and output (4.4 L/min) but fails peak pressure (186 mmHg). The model that reports these numbers can prescribe output and mean pressure; a closed loop cannot.
Compartment swings. The arteriolar volume swing is 23% subendocardially against the 17% Mynard's model gives, which fails, and 3% subepicardially against its 1.6%. The venular swing fails by far more: 65% against 7 subendocardially and 18% against 2.3 subepicardially. Those are that paper's model output, not measurements. The arteriolar diameter swing passes, 12% against Munneke's 12 and the 10 to 24% measured as Munneke reports. The subendocardial venular diameter swing, 41% against Munneke's 10 to 18, fails.
Read in full text:
The originating description of the supply:demand index, read in full text: Buckberg GD, Fixler DE, Archie JP, Hoffman JIE. Experimental subendocardial ischemia in dogs with normal coronary arteries. Circ Res 1972;30(1):67–81. doi:10.1161/01.res.30.1.67. It defines DPTI as used here, measured with mean atrial pressure in place of LV diastolic pressure. Its denominator is the tension time index, from the aortic systolic pressure curve, which it says is more accurately a systolic pressure time index. SPTI from LV pressure, as used here, is its authors' 2014 wording.
Paediatric and congenital values for the index, all read as abstract only and quoted here as single-centre catheter cohorts, not as normal limits. Invasive, by catheter:
Non-invasive tonometric values in adolescents, on the other scale, are 132 ± 22% at normal weight and 114 ± 26% in overweight peers (Tocci 2021, above; its abstract states the opposite direction to its own Table 2, and the table is used here).
Read as an abstract or less:
We found no published paediatric normal range for this index by age or body size, no neonatal or preterm value, no value in a child with no cardiac lesion at all, and no paediatric threshold validated against ischaemia.
The model runs in this browser; nothing you set on this page is stored or sent.
Educational use only. This is a teaching model of a mechanism, not a patient-specific prediction and not a treatment recommendation; check every number against the primary sources before relying on it.