Cardiogenic Shock and Ventricular Pump Failure

Cardiogenic shock develops when cardiac performance becomes a major limitation to sustaining organ perfusion. The failing circulation combines inadequate forward delivery with changing filling pressures, vascular responses, and metabolic demands. Understanding it requires more than identifying a low ejection fraction: the relevant question is how the heart, its loads, and the connected organs are functioning together.

72.1 Pump failure is a circulatory mechanism, not an ejection fraction

The heart must receive blood, fill at tolerable pressures, generate force, eject useful forward flow, and maintain a rhythm that coordinates these tasks. Failure at any of these steps can limit the circulation. Cardiogenic shock describes the resulting perfusion failure when cardiac dysfunction is central to the process, rather than one echocardiographic appearance.

Acute coronary occlusion, myocarditis, decompensated chronic myocardial disease, severe valve dysfunction, and arrhythmias can initiate different forms of pump failure. A sudden lesion encounters different chamber adaptation and reserve from a longstanding cardiomyopathy. The same measured output can therefore have different implications in two patients.

A reduced ejection fraction identifies a smaller fraction of end-diastolic chamber volume leaving during systole. It does not state the absolute forward volume, the pressure generated, the adequacy of filling, or tissue requirements. Some patients tolerate chronically reduced fractions without shock; others develop shock despite a relatively preserved fraction.

Arterial pressure is similarly incomplete. Compensatory vasoconstriction can maintain a pressure while forward flow is severely restricted. Conversely, vasodilation can produce profound hypotension with less severe myocardial impairment. The presence of myocardial dysfunction does not establish that it is the only cause of the current shock.

The phenotype also changes over time. Ischemia, inflammation, hypoxemia, acidosis, treatment, and congestion can add new limitations to the original lesion. An initially predominantly cardiac problem can become a mixed circulatory and metabolic syndrome without ceasing to contain important pump failure.

The physiological assessment therefore connects cause with performance and consequences. Forward flow, pressure, oxygen content, ventricular filling, valve behavior, rhythm, regional perfusion, and organ trajectory should support a coherent explanation. Shock cannot be diagnosed or excluded from a single fraction, pressure, or lactate concentration in isolation.

72.2 Systolic failure depends on the load the ventricle encounters

During systole, the ventricle generates pressure and shortens against its arterial and valvular load. Reduced systolic capability leaves more blood behind when the same load must be overcome. Alternatively, an increased load can reduce ejection even without a primary deterioration in contractile activation.

End-systolic elastance describes a modeled relationship between end-systolic pressure and volume. Effective arterial elastance summarizes aspects of arterial loading through a pressure-to-stroke-volume relationship. Their interaction helps explain why ventricular performance cannot be interpreted independently of load, although neither index captures every property of an unstable human circulation.

Sunagawa and colleagues tested a ventricular–arterial framework in isolated canine ventricles. The experiments demonstrated how ventricular properties, arterial impedance, and filling jointly determine stroke volume. This mechanistic preparation supports the coupling concept; it does not establish one optimal bedside coupling ratio for cardiogenic shock. 1

Define end-diastolic and end-systolic volumes in mL, stroke volume as SV in mL/beat, heart rate as HR in beats/min, and cardiac output as CO in L/min. With no important regurgitation or shunt affecting the relationship:

For constructed volumes of 150 and 110 mL, SV is 40 mL. At 90 beats/min, output is 90 × 40/1000 = 3.6 L/min. The chamber ejection fraction is 40/150, or approximately 26.7%. Whether this output meets tissue needs depends on body size, oxygen content, distribution, and demand; the example supplies no shock threshold.

Dilation may initially support ejection through additional filling, but also increases wall stress at a given transmural pressure and wall thickness. A weak, enlarged ventricle can consequently spend substantial energy generating pressure while delivering little useful flow. Lowering its load may help ejection, while excessive reduction of arterial pressure can compromise coronary and systemic perfusion. Those competing effects require observation rather than a universal pressure rule.

72.3 Filling pressure can rise without providing useful preload

The failing ventricle still requires sufficient filling. The difficulty is determining whether additional inflow produces useful stroke volume or mainly increases pressure upstream. A high measured filling pressure is not proof that the myocardium has an effective preload reserve, and a low output is not proof that more volume will help.

Diastolic relaxation and chamber stiffness influence the pressure needed to accommodate blood. A stiff ventricle can require a higher filling pressure to achieve a given volume. Its steep diastolic pressure–volume relationship means that a small volume addition can produce a large pressure rise; it does not imply a proportionately large volume increase for every pressure change.

External constraints matter as well. Pleural and pericardial pressures contribute to intracavitary measurements. Right ventricular enlargement can restrict left ventricular filling through the septum and pericardium. A pulmonary artery wedge pressure therefore requires its measurement conditions and cannot automatically be treated as left ventricular transmural pressure or fiber stretch.

Upstream pressure has consequences regardless of whether it improves useful filling. Increased pulmonary microvascular hydrostatic pressure favors fluid filtration, while lung permeability, lymphatic drainage, and time determine the resulting edema. Systemic venous hypertension can impair organ drainage and increase tissue pressure even when arterial inflow has also fallen.

Mullens and colleagues associated greater venous congestion with worsening renal function in advanced decompensated heart failure. This human cohort supports attention to downstream pressure alongside output, but neither its association nor its achieved pressures define a universal decongestion target in cardiogenic shock. 2

Congestion and reduced useful circulating filling can coexist because blood and extracellular fluid are distributed unevenly. A flow response to additional filling, when demonstrable, also does not establish that further fluid is tolerable. The balance depends on the incremental forward benefit and the pulmonary, venous, and ventricular costs.

72.4 Useful forward ejection differs from total chamber emptying

A ventricle can empty a substantial proportion of its cavity without delivering an adequate volume to the systemic circulation. Regurgitation and intracardiac shunting break the simple equivalence between total chamber stroke volume and useful forward output.

In mitral regurgitation, left ventricular ejection divides between the aorta and the left atrium. The regurgitant pathway changes the load against which the chamber empties. A relatively preserved ejection fraction can consequently coexist with low forward flow and marked pulmonary venous pressure.

For an explicitly constructed example, let end-diastolic volume be 120 mL and end-systolic volume 48 mL. Total chamber stroke volume is 72 mL and ejection fraction 60%. If 42 mL returns to the left atrium during each beat, only 30 mL goes forward. At 100 beats/min, forward output is 3.0 L/min, rather than the 7.2 L/min that would be inferred by treating all chamber emptying as systemic delivery.

These assumed volumes are conservation arithmetic, not a method for estimating an individual regurgitant fraction from visual inspection. Quantification requires appropriate acquisition and accounting for all relevant pathways. Acute regurgitation can also raise atrial pressure sharply before the atrium has adapted.

The SHOCK registry’s severe post-infarction mitral-regurgitation cohort had higher average ejection fraction than its predominantly left ventricular failure cohort despite similarly poor outcomes. The observational comparison illustrates the limitation of fraction alone; treatment comparisons within that registry cannot establish a randomized surgical benefit. 3

Valve obstruction creates another mismatch: the ventricle may generate pressure well above the arterial pressure measured beyond a narrowed outlet. Septal rupture can redirect flow through the pulmonary circuit. Rhythm disturbances alter filling intervals, atrial contribution, synchrony, and the number of effective contractions. A fast rhythm can be compensatory or itself contribute to failure. Each mechanism requires identifying the actual flow pathway rather than labeling every low-output state as inadequate contractility.

72.5 The ventricles fail as a connected system

The two ventricles are connected in series: right ventricular output crosses the lungs and becomes left ventricular inflow. A poorly supplied left ventricle can therefore eject little despite preserved intrinsic systolic capability. Conversely, elevated left atrial pressure can increase pulmonary vascular loading and burden the right ventricle.

Shared structures add direct interaction. Right ventricular dilation can displace the septum and raise pericardial constraint, reducing the space and pressure gradient available for left ventricular filling. Systemic congestion can thus coexist with a relatively small left ventricular cavity. That combination should not automatically be called simple volume depletion.

Brinker and colleagues demonstrated altered septal configuration during acute right ventricular loading in healthy volunteers. This human physiological observation supports mechanical interdependence, while the respiratory maneuver and healthy population do not reproduce every feature of clinical biventricular shock. 4

Right ventricular myocardial ischemia or dysfunction can also be a primary pump lesion. Detailed pressure overload, pulmonary obstruction, and obstructive-shock mechanisms are developed in Chapter 73. Here the essential point is that left-sided support cannot be interpreted without checking whether the right heart can deliver blood through the lungs.

Respiratory conditions modify both sides. Positive intrathoracic pressure can reduce left ventricular transmural load while reducing venous return. Lung inflation can either relieve hypoxic vascular constriction through recruitment or increase pulmonary vascular loading through overdistension. Strong spontaneous effort can increase left ventricular transmural load and the work required from the circulation.

The net response depends on recruitability, chest-wall mechanics, pulmonary vascular conditions, volume distribution, and ventricular reserve. A lower wedge pressure after a respiratory change may reflect altered transmitted pressure as well as a true change in loading. Improved oxygenation does not independently establish increased cardiac output. Both ventricular function and the pressure reference belong in the interpretation.

72.6 Coronary supply and myocardial demand create a failure loop

The failing heart must supply its own working muscle through the coronary circulation. Reduced arterial pressure can restrict this supply, while elevated ventricular diastolic pressure and tissue compression oppose inflow. The left ventricular subendocardium is especially dependent on the pressure and time available during diastole.

Tachycardia increases the number of contractions and can shorten diastolic perfusion time. Dilation increases wall stress, while stronger adrenergic activation increases contractile and calcium-handling work. Poor systemic output therefore does not imply that the heart has a low oxygen requirement.

Buckberg and colleagues’ canine experiments produced subendocardial ischemia without an anatomical coronary obstruction by changing the pressure and timing environment. This establishes a physiological route to supply limitation; no experimental pressure–time ratio becomes a universal clinical ischemia threshold. 5

The resulting loop can reinforce itself: impaired ejection increases residual volume and filling pressure; greater stress and less favorable perfusion promote ischemia; ischemia impairs contraction and relaxation further. Anemia, hypoxemia, coronary stenosis, and excessive respiratory or metabolic work can tighten the constraint.

An acute coronary lesion is a different and potentially concurrent mechanism. Troponin elevation can indicate myocardial injury without establishing infarction. An ischemic explanation and appropriate evidence remain necessary for the infarction diagnosis. A systemic stressor does not by itself determine the coronary diagnosis; Chapter 13 develops the current distinction.

The original SHOCK trial randomized patients with infarct-related left ventricular failure to early revascularization or initial medical stabilization. Its primary 30-day mortality difference was not statistically significant, while six-month mortality, a secondary endpoint, was lower with early revascularization. The trial distinguishes treatment of an initiating coronary lesion from merely changing a pressure or flow surrogate. It should not be rewritten as an immediate universal reversal of pump failure or as evidence about nonischemic shock. 6

72.7 Vascular responses can compensate or create mixed shock

Reduced cardiac output normally stimulates sympathetic vasoconstriction and changes venous capacity. Arteriolar constriction helps sustain arterial pressure, while venoconstriction can mobilize blood toward the central circulation. These are compensatory responses, but their benefit depends on the heart’s ability to handle the resulting load and filling.

Greater arterial resistance can make ejection more difficult for a weak ventricle. Mobilized venous blood can raise filling pressure without a matching increase in useful output. The same response that preserves pressure may consequently aggravate myocardial stress or congestion. This is a tradeoff within an integrated feedback system, not proof that compensation is intrinsically harmful.

Vasodilation can also accompany cardiac shock. Ischemic tissue injury, inflammatory signaling, infection, medications, and prolonged critical illness can reduce vascular responsiveness. The circulation may then contain both inadequate pump performance and insufficient vascular tone. Detailed inflammatory and vascular mechanisms are developed in Chapters 66 and 71.

Kohsaka and colleagues examined severe systemic inflammatory features within the SHOCK population. Lower vascular resistance near shock onset and later inflammatory or infectious patterns were associated, with important timing and selection limits. These observations support a mixed phenotype; they do not prove that one cytokine or gut-translocation mechanism caused every episode of post-infarction vasodilation. 7

Temperature, skin appearance, and pulse pressure can change as the vascular component evolves. A warm extremity does not exclude low useful output, and a cold extremity can reflect vasoconstriction without describing all visceral perfusion. Prior drugs further change the observed phenotype.

Calculated vascular resistance also requires caution: it is a pressure difference divided by output. It can rise because output fell rather than because an independently measured constrictor process intensified. Serial interpretation should inspect its component measurements, current treatment, and evidence from the ventricles and tissues before assigning the dominant limitation.

72.8 Oxygen delivery and microvascular access determine the organ consequence

Forward blood flow matters because it transports oxygen and substrates. The oxygen consequence depends on arterial content as well as volume flow. A high arterial saturation cannot compensate for severe anemia or inadequate output by itself.

Let DO₂ be systemic oxygen delivery in mL O₂/min, CO forward output in L/min, and CaO₂ arterial oxygen content in mL O₂/dL. The conversion between liters and deciliters gives:

At the constructed output of 3.6 L/min and content 16 mL/dL, delivery is 3.6 × 10 × 16 = 576 mL O₂/min. This describes global incoming oxygen, not oxygen consumption or a universal requirement. Temperature, muscle activity, respiratory work, inflammation, and cellular state change the demand it must support.

Tissues can initially increase extraction as delivery falls. A lower mixed venous oxygen content may reflect that compensation, but the available reserve differs among organs. A central venous sample also has different regional weighting from mixed pulmonary arterial blood and cannot be substituted without qualification.

Microvascular access adds another limitation. De Backer and colleagues observed sublingual small-vessel perfusion abnormalities in severe heart failure and cardiogenic shock, with more marked abnormalities among nonsurvivors. The study establishes an accessible regional observation and outcome association, not a map of every organ or proof that a particular microvascular intervention improves survival. 8

Venous congestion can further impair tissue exchange and drainage while total output is low. Lactate may rise from supply limitation, adrenergic glycolysis, altered disposal, or several processes together. A normal value does not exclude regional inadequacy, and a falling value does not certify restored microvascular distribution.

Thus, pressure, output, arterial content, extraction, congestion, and organ function contribute different evidence. Their agreement strengthens an interpretation; their disagreement identifies a physiological question rather than a reason to force all observations toward one global delivery target.

72.9 Measurements describe performance under the current conditions

Echocardiography identifies chamber size, ventricular function, valve lesions, septal behavior, and some consequences of congestion. Forward-flow estimation requires an appropriate outflow area and velocity integral, with attention to alignment and beat selection. Apparent visual improvement in contraction is not a quantitative measurement of greater systemic output.

Invasive pressures add information when correctly referenced and acquired. Right atrial, pulmonary arterial, and wedge measurements describe different locations. Respiratory pressure, valve lesions, catheter position, and waveform interpretation affect their meaning. Thermodilution and Fick estimates have their own sampling and methodological limits, especially during rapidly changing conditions or mechanical support.

Cardiac power combines pressure with flow. The conventional reported cardiac power output, CPO, uses mean arterial pressure, MAP, in mmHg and CO in L/min; approximately 451 converts their product into watts:

This is a simplified mean hydraulic index referenced through arterial pressure. It is not total myocardial energy expenditure or a complete pressure–volume description. An alternative using an arterial-to-right-atrial pressure difference answers a somewhat different question; conventions should not be silently mixed.

Fincke and colleagues found conventional cardiac power strongly associated with mortality in the SHOCK registry. Missing measurements, patient selection, and concurrent support limit interpretation. A prognostic association does not demonstrate that targeting one power value improves outcome. 9

For constructed MAP 60 mmHg and CO 3.6 L/min, CPO is 216/451, approximately 0.479 W. If MAP becomes 75 while CO falls to 2.8, CPO is 210/451, approximately 0.466 W. Pressure has increased while this mean hydraulic index has slightly fallen. Neither value selects treatment or certifies local perfusion.

Derived indices are not independent additional measurements. Output appears within cardiac index, resistance, oxygen delivery, and power. Concordant changes in those calculations may largely repeat one underlying flow observation. Reliable timing and the separate raw quantities remain essential.

72.10 Drug effects must be separated from outcomes

Inotropic stimulation can increase activation, calcium cycling, and ejection. It can also increase heart rate, arrhythmia susceptibility, and myocardial oxygen requirements. Vasoconstriction can support perfusion pressure while changing arterial load and venous capacity. Drugs described by one dominant action usually affect several parts of this system.

An inodilator can improve ejection or pulmonary vascular conditions while lowering systemic pressure. A pressor can raise pressure with a variable effect on stroke volume. The observed result depends on receptor responsiveness, drug exposure, rhythm, loading, renal handling, and which mechanism currently limits the circulation.

OptimaCC randomized 57 patients with post-infarction cardiogenic shock to epinephrine or norepinephrine. Cardiac-index evolution was similar, but refractory shock was more frequent with epinephrine, prompting early termination. Greater tachycardia and lactic acidosis also occurred with epinephrine. The small, selected trial distinguishes comparable pressure/flow effects from safety and metabolic effects; it does not establish that every lactate increase represents worsening oxygen delivery or provide a universal pressor sequence. 10

DOREMI randomized 192 patients with cardiogenic shock to milrinone or dobutamine in a blinded comparison. The primary in-hospital composite outcome and major individual outcomes did not differ significantly. This was not a proof of exact equivalence, nor does it make the drugs’ pharmacology interchangeable in every rhythm, pressure, or renal state. 11

Serial response should therefore connect the intended effect with possible costs. A higher pressure alongside falling forward flow differs from greater flow with stable tissue perfusion. A lower measured filling pressure may reflect altered surrounding pressure rather than reduced congestion; the organ response and measurement context help distinguish them.

Changing a physiological variable is valuable when it addresses a limitation, but it is not itself proof of improved survival or recovery. The relevant endpoint is the net functioning of the circulation and organs, alongside adverse effects and the underlying disease trajectory.

72.11 Mechanical support changes flow pathways and ventricular loads

Temporary support devices move blood through different routes. A transvalvular left ventricular pump transfers blood from the ventricle into the aorta, changing chamber loading while contributing systemic flow. It still depends on adequate blood reaching its inlet. Displayed pump flow cannot simply be added to an unchanged presupport native output because native ejection and recirculating pathways also change.

Peripheral venoarterial extracorporeal support drains venous blood and returns oxygenated blood to the arterial circulation. It can support systemic delivery while increasing the arterial load encountered by the left ventricle. Residual pulmonary inflow can then accumulate when the ventricle cannot eject effectively. Oxygen distribution also depends on mixing between extracorporeal return and native output.

Donker and colleagues’ closed-loop simulation demonstrated that systemic hemodynamic improvement during venoarterial support can coexist with left ventricular overload. The modeled effects depended on loading and adjunct interventions; its numerical responses are not fixed patient corrections or proven clinical unloading strategies. 12

Counterpulsation alters the timing of aortic pressure rather than functioning as the same continuous flow pathway. IABP-SHOCK II randomized 600 patients with infarct-related shock and planned early revascularization; routine balloon support did not significantly reduce 30-day mortality. That result concerns the tested strategy, not every mechanical complication or another device. 13

DanGer Shock randomized 360 patients with ST-elevation myocardial infarction and shock, analyzing 355. The microaxial-pump strategy reduced 180-day mortality, while increasing its composite safety events and use of renal replacement therapy. Eligibility excluded predominant right ventricular failure, post-infarction mechanical complications, and persistent coma after out-of-hospital arrest. ECLS-SHOCK randomized 420 patients with infarct-related shock and planned early revascularization, analyzing 417. Routine early extracorporeal support did not reduce 30-day mortality and increased bleeding and vascular complications. Different devices, populations, strategies, and follow-up periods prevent treating these as a direct head-to-head comparison or generalizing either result to every pump-failure phenotype. 14 15

The physiological question remains what the device unloads, what it loads, where useful blood and oxygen go, and whether the patient can recover or transition beyond that support. A better displayed flow does not establish myocardial recovery, adequate regional delivery, or favorable net outcome.

72.12 Integrate pressure, forward flow, congestion, and trajectory

Return to the constructed circulation with MAP 60 mmHg, forward CO 3.6 L/min, and arterial oxygen content 16 mL/dL. Delivery is 576 mL O₂/min and conventional cardiac power approximately 0.479 W. Suppose a later observation shows MAP 75, CO 2.8, and unchanged content. Delivery becomes 2.8 × 10 × 16 = 448 mL O₂/min, while power becomes approximately 0.466 W.

Delivery has fallen by 128 mL O₂/min, or approximately 22.2%, despite the higher pressure. The pressure rise could still benefit a pressure-limited territory. These arithmetic relationships do not establish whether the change arose from arterial loading, reduced filling, ischemia, rhythm, right ventricular limitation, or measurement error.

Now suppose right atrial and wedge pressures rise while respiratory distress increases. Those observations raise concern about pressure accumulation and cardiopulmonary interaction, but their references, respiratory conditions, and timing must be checked. A visual ejection fraction cannot resolve forward volume if regurgitation or an outflow lesion is present.

If lactate also rises, it adds evidence of a changing production–disposal balance. Adrenergic exposure, hepatic handling, oxygen delivery, and regional perfusion may contribute together. A later fall would likewise require interpretation rather than being accepted as proof that every organ has recovered.

The useful synthesis identifies the primary lesion and the current limiting relationships: filling into useful ejection, capability against load, coronary supply against myocardial work, pulmonary delivery between the ventricles, and systemic delivery against tissue need. Measurements should be repeated in comparable conditions, with attention to whether an intervention improved its intended limitation and introduced another.

Cardiogenic shock is therefore understood through a trajectory rather than a single low number. Pressure, forward flow, congestion, oxygen balance, rhythm, and organ function must be connected. This preserves the physiological reasons for support while recognizing that the benefit of a particular strategy requires its own clinical evidence.

References

1. Sunagawa K, Maughan WL, Burkhoff D, Sagawa K. Left ventricular interaction with arterial load studied in isolated canine ventricle. Am J Physiol. 1983;245:H773–H780. Original experiment.

2. Mullens W, Abrahams Z, Francis GS, et al. Importance of venous congestion for worsening of renal function in advanced decompensated heart failure. J Am Coll Cardiol. 2009;53:589–596. Original cohort.

3. Thompson CR, Buller CE, Sleeper LA, et al. Cardiogenic shock due to acute severe mitral regurgitation complicating acute myocardial infarction: a report from the SHOCK Trial Registry. J Am Coll Cardiol. 2000;36:1104–1109. Original registry study.

4. Brinker JA, Weiss JL, Lappé DL, et al. Leftward septal displacement during right ventricular loading in man. Circulation. 1980;61:626–633. Original human experiment.

5. Buckberg GD, Fixler DE, Archie JP, Hoffman JI. Experimental subendocardial ischemia in dogs with normal coronary arteries. Circ Res. 1972;30:67–81. Original experiment.

6. Hochman JS, Sleeper LA, Webb JG, et al. Early revascularization in acute myocardial infarction complicated by cardiogenic shock. N Engl J Med. 1999;341:625–634. Original SHOCK trial.

7. Kohsaka S, Menon V, Lowe AM, et al. Systemic inflammatory response syndrome after acute myocardial infarction complicated by cardiogenic shock. Arch Intern Med. 2005;165:1643–1650. Original SHOCK cohort analysis.

8. De Backer D, Creteur J, Dubois MJ, Sakr Y, Vincent JL. Microvascular alterations in patients with acute severe heart failure and cardiogenic shock. Am Heart J. 2004;147:91–99. Original microvascular study.

9. Fincke R, Hochman JS, Lowe AM, et al. Cardiac power is the strongest hemodynamic correlate of mortality in cardiogenic shock: a report from the SHOCK trial registry. J Am Coll Cardiol. 2004;44:340–348. Original registry analysis.

10. Levy B, Clere-Jehl R, Legras A, et al. Epinephrine versus norepinephrine for cardiogenic shock after acute myocardial infarction. J Am Coll Cardiol. 2018;72:173–182. Original OptimaCC trial.

11. Mathew R, Di Santo P, Jung RG, et al. Milrinone as compared with dobutamine in the treatment of cardiogenic shock. N Engl J Med. 2021;385:516–525. Original DOREMI trial.

12. Donker DW, Brodie D, Henriques JPS, Broomé M. Left ventricular unloading during veno-arterial ECMO: a simulation study. ASAIO J. 2019;65:11–20. Original model study.

13. Thiele H, Zeymer U, Neumann FJ, et al. Intraaortic balloon support for myocardial infarction with cardiogenic shock. N Engl J Med. 2012;367:1287–1296. Original IABP-SHOCK II trial.

14. Møller JE, Engstrøm T, Jensen LO, et al. Microaxial flow pump or standard care in infarct-related cardiogenic shock. N Engl J Med. 2024;390:1382–1393. Original DanGer Shock trial; registered eligibility.

15. Thiele H, Zeymer U, Akin I, et al. Extracorporeal life support in infarct-related cardiogenic shock. N Engl J Med. 2023;389:1286–1297. Original ECLS-SHOCK trial.