Interpreting end-tidal CO2 (EtCO2) waveforms during anesthesia — pattern recognition, airway safety, and CPR quality feedback
Capnography is the continuous, real-time measurement and graphic display of the partial pressure of carbon dioxide in exhaled gas. Unlike pulse oximetry, which reports oxygenation, capnography reports ventilation breath-by-breath — and it does so faster than any other monitor in the operating room. A normal capnogram is not a single number; it is a reproducible four-phase waveform, and understanding its shape is the foundation for recognizing every abnormality that follows.
Carbon dioxide is produced continuously by cellular aerobic metabolism (the Krebs cycle) at a rate of roughly 200 mL/min in a resting adult, rising several-fold with exercise, fever, sepsis, or malignant hyperthermia. This metabolic CO2 diffuses from tissue into capillary blood, is transported to the lungs mostly as bicarbonate (via carbonic anhydrase in red cells) and partly bound to hemoglobin, and diffuses across the alveolar-capillary membrane down its partial-pressure gradient into alveolar gas — a process roughly 20 times more efficient by diffusion coefficient than oxygen exchange.
The amount of CO2 that ultimately appears in exhaled gas depends on three linked variables: production (metabolism), transport (cardiac output and blood CO2 content), and elimination (alveolar ventilation). Because capnography samples the very last portion of each exhaled breath — the gas that has most fully equilibrated with alveolar gas — end-tidal CO2 (EtCO2) approximates alveolar PCO2, which in healthy lungs with well-matched ventilation and perfusion runs only 2–5 mmHg below arterial PaCO2. This is why capnography is used clinically as a rapid, non-invasive proxy for arterial CO2 and, indirectly, for both ventilation and pulmonary blood flow.
Every normal breath traces the same characteristic waveform, conventionally divided into four phases:
Phase I — Inspiratory baseline: the earliest portion of exhalation, consisting of CO2-free gas from the anatomic dead space (conducting airways: trachea, bronchi) that never reached alveoli. The trace is flat at zero.
Phase II — Expiratory upstroke: as alveolar gas rich in CO2 begins mixing with the residual dead-space gas, the trace rises sharply — normally in well under a second. The steepness of this upstroke reflects how quickly and uniformly alveolar gas reaches the airway opening.
Phase III — Alveolar plateau: once alveolar gas dominates the sampled stream, the trace flattens into a near-horizontal plateau with only a gentle upward slope, reflecting continued, fairly uniform emptying of alveolar units with slightly varying time constants. The value at the very end of Phase III, just before inspiration begins, is the EtCO2 — the number displayed and trended by the monitor.
Phase IV — Inspiratory downstroke: fresh, CO2-free inspired gas enters the airway and displaces the sampled gas, and the trace falls sharply back to the zero baseline, beginning the next cycle at Phase I.
A normal waveform is therefore a rectangular, box-like trace: flat, sharp up, gently sloped plateau, sharp down — repeating breath after breath with remarkable consistency in a healthy, mechanically ventilated patient.
A systematic approach to any capnogram should proceed in order: (1) Rate — count breaths per minute from the waveform itself. (2) Baseline — does Phase I return fully to zero? (3) Shape — is the upstroke sharp and the plateau flat, or sloped? (4) Height — what is the EtCO2 value at end-plateau? (5) Trend — is EtCO2 stable, rising, falling, or suddenly absent across successive breaths? This five-step sequence catches the overwhelming majority of clinically important abnormalities before a specific diagnosis is even considered.
Continuous capnography during general anesthesia is a core component of the ASA (American Society of Anesthesiologists) Standards for Basic Anesthetic Monitoring, required for every patient receiving general anesthesia and strongly recommended for moderate/deep sedation. Its mandate rests on a simple fact: capnography is the single fastest and most specific bedside confirmation that a patient is being ventilated at all, breath by breath, in real time — faster than pulse oximetry, which lags oxygen desaturation by tens of seconds to minutes, especially in preoxygenated patients.
The waveform, not just the number, is the monitor. A normal number with an abnormal shape can be as informative as an abnormal number: a "normal" EtCO2 of 38 mmHg displayed during a period of no actual waveform (frozen or extrapolated by some devices) is not reassuring. This is why every modern capnography display shows the continuous scrolling trace, not merely a digital readout.
Of all the applications of capnography, none is more immediately life-saving than confirmation of correct endotracheal tube placement. Absence of a capnogram after intubation — or a capnogram that appears briefly and then rapidly decays to nothing — is the most reliable sign that the tube has been placed in the esophagus rather than the trachea, and it must trigger immediate reassessment before any other clinical sign is trusted.
Traditional confirmation methods for endotracheal tube placement — auscultation of bilateral breath sounds, absence of epigastric sounds, visualization of chest rise, tube fogging — are all subject to false reassurance. Breath sounds can transmit through the mediastinum from an esophageal tube; chest rise can be produced by gastric insufflation; fogging occurs with humidified air regardless of location. Capnography avoids all of these pitfalls because it detects the one thing that only the lungs can produce during ventilation: a rhythmic, physiologic CO2 waveform derived from pulmonary gas exchange.
When a tube is correctly placed in the trachea, each positive-pressure breath delivered by bag or ventilator produces the full four-phase waveform described in Stage 1, essentially from the very first breath. When a tube is in the esophagus, there is ordinarily no true alveolar gas exchange occurring through it — any capnographic signal detected is either absent entirely, or a small, low, rapidly diminishing signal from CO2 that was present in the stomach from previously swallowed air or brief mask ventilation. That residual gastric CO2 washes out within a handful of breaths, and the waveform decays toward a flat line.
The clinical rule of thumb: a sustained, normal-shaped capnogram present for at least six consecutive breaths is considered definitive confirmation of tracheal placement. A waveform that is absent from the outset, or present-but-rapidly-decaying over just a few breaths, should be treated as esophageal intubation until proven otherwise — pull the tube, reoxygenate, and reattempt.
A flat or absent capnogram is not always esophageal intubation — the differential must be worked through quickly:
• Esophageal intubation: no or rapidly decaying waveform from the first breaths, typically with no chest rise on careful inspection and gastric distension with repeated bagging.
• Complete airway obstruction / mainstem occlusion: no waveform, but often with high resistance felt on the bag and no chest movement at all.
• Ventilator/circuit disconnection: sudden flat line with an obvious mechanical disconnect, apnea alarm, and no delivered breath at all.
• Cardiac arrest with correctly placed tube: a waveform is usually still present but of markedly reduced amplitude (covered in Stage 5), because ventilation continues even though pulmonary blood flow has collapsed — this is different from true esophageal placement where the tube itself is not sampling alveolar gas.
• Sensor or sampling line fault: a flat line accompanied by no other clinical abnormality and resolved by checking connections — always exclude equipment failure before invasive re-intervention when the patient otherwise appears well.
When airway resistance rises — from bronchospasm, COPD, a kinked or partially obstructed endotracheal tube, or a mucus plug — the capnogram loses its crisp rectangular shape. The sharp Phase II upstroke becomes a slow, sloped ramp, and the flat Phase III plateau is replaced by a continuously rising ramp, producing the unmistakable triangular "shark-fin" silhouette that anesthesiologists learn to recognize at a glance.
The sharp Phase II upstroke of a normal capnogram exists because essentially all alveolar units empty at nearly the same rate, so CO2-rich alveolar gas floods the airway almost simultaneously. Airway obstruction — from bronchospasm-induced smooth muscle constriction, mucosal edema, secretions, or a mechanically narrowed endotracheal tube — introduces wide variability in regional time constants (time constant = airway resistance × alveolar compliance). Some lung units, with low resistance, empty quickly and contribute early CO2; others, with high resistance, empty slowly and continue contributing CO2 well into what would normally be the flat plateau.
The net effect on the waveform is twofold: the upstroke (Phase II) becomes progressively less vertical as emptying is staggered over a longer interval, and the plateau (Phase III) never truly flattens — it keeps climbing throughout expiration because slow-emptying units keep adding CO2-rich gas right up until inspiration begins. The resulting shape resembles a shark's dorsal fin: a sloped ramp rising continuously from baseline to peak, with no discrete corner between "upstroke" and "plateau." The severity of the slope correlates directly with the severity of airway obstruction, making the shark-fin pattern a useful non-invasive severity indicator that can be trended breath-to-breath during treatment.
Because the shark-fin pattern reflects mechanical airway resistance rather than a single anatomic lesion, it is seen with bronchospasm (asthma, anaphylaxis, light anesthesia with airway irritation), COPD/emphysema with dynamic airway collapse, retained secretions or mucus plugging, and simple mechanical causes such as a kinked, bitten, or partially obstructed endotracheal tube — always inspect the tube and circuit before assuming a purely pharmacologic bronchospasm.
Recognizing the shark-fin pattern intraoperatively should prompt a rapid, staged response: first exclude and correct mechanical causes (kinked tube, mainstem intubation, secretions needing suction); then deepen the volatile anesthetic, which itself provides bronchodilation; then administer inhaled or intravenous β2-agonists (albuterol) and, for severe or refractory bronchospasm, consider epinephrine and corticosteroids while ruling out anaphylaxis as the trigger.
Capnography provides continuous, breath-by-breath feedback on treatment efficacy that precedes any change in oxygen saturation: as bronchodilator therapy takes effect, the sloped upstroke progressively steepens and the plateau progressively flattens back toward the normal rectangular shape, often within one to two minutes of effective treatment — long before spirometric or auscultatory improvement would be appreciated. Concurrently rising peak inspiratory pressures on the ventilator and a prolonged expiratory phase corroborate the capnographic diagnosis.
Two distinct but often co-occurring problems produce a rising EtCO2 trend: rebreathing of exhaled CO2 due to a circuit fault, and true alveolar hypoventilation from inadequate minute ventilation. The capnogram distinguishes them by where the abnormality appears — a baseline that never returns to zero implicates the breathing circuit, while a normal baseline with a climbing plateau height implicates ventilation itself.
In a properly functioning circle breathing circuit, unidirectional valves and CO2-absorbing granules (soda lime or a similar absorbent) ensure that every breath the patient inspires is free of CO2, so Phase I of the capnogram sits flat at zero. Rebreathing occurs when previously exhaled, CO2-containing gas is re-inspired without adequate scrubbing — most commonly from exhausted or channeled CO2 absorbent (the granules turn color when spent, but channeling can cause localized exhaustion before the color change is obvious), an incompetent or stuck inspiratory or expiratory unidirectional valve, or, in non-rebreathing/Mapleson circuits, an inadequate fresh gas flow relative to minute ventilation.
The capnographic signature is a baseline that fails to return fully to zero between breaths — the entire waveform appears to float upward off the axis. This is a distinct pattern from hypoventilation: it is a circuit/equipment fault, not a patient physiology problem, and the fix is mechanical (replace absorbent, check/replace faulty valves, increase fresh gas flow) rather than a change in ventilator settings.
A rapidly rising baseline mid-case, especially with a fresh canister of absorbent still showing purple/violet exhaustion color change, should prompt immediate inspection of the inspiratory and expiratory unidirectional valves — a valve stuck open allows retrograde flow of exhaled gas back into the inspiratory limb regardless of absorbent condition.
True alveolar hypoventilation — inadequate minute ventilation relative to metabolic CO2 production — produces a different capnographic fingerprint: the baseline (Phase I) remains at zero, but the plateau height (Phase III, and therefore EtCO2) climbs progressively over successive breaths. Causes include inadequate tidal volume or respiratory rate settings, residual neuromuscular blockade with weak spontaneous effort, opioid-induced respiratory depression, or increased CO2 production outstanding a fixed ventilation (fever, malignant hyperthermia, laparoscopic CO2 insufflation being absorbed systemically, sepsis, thyroid storm).
Because EtCO2 tracks arterial PaCO2 closely in healthy lungs, a steadily climbing EtCO2 trend over several minutes is a sensitive early warning of inadequate ventilation — often detectable well before oxygen saturation falls, particularly in a preoxygenated, well-oxygenated patient whose SpO2 can remain normal for many minutes despite significant hypercapnia. Management is directed at the underlying cause: increasing tidal volume or rate, reversing residual paralysis, reducing opioid effect, or — if CO2 production itself has risen sharply and unexpectedly — considering malignant hyperthermia, a life-threatening diagnosis for which an abrupt, otherwise-unexplained EtCO2 surge is one of the earliest and most sensitive clinical clues.
Perhaps the most dramatic and prognostically important capnographic event is a sudden, severe fall in EtCO2 while ventilation continues completely unchanged. Because EtCO2 depends on pulmonary blood flow delivering CO2 to the alveoli as much as it depends on ventilation, an abrupt drop with stable ventilation points directly at the circulation — cardiac arrest, massive pulmonary embolism, or profound hypotension/shock — and capnography becomes an indispensable, real-time guide to resuscitation quality.
EtCO2 is the product of two independent physiologic processes arriving together at the alveolus: ventilation must bring alveolar gas to the airway opening, and perfusion must deliver CO2-laden venous blood to the alveolar capillaries in the first place. Stages 1 through 4 concern derangements of the ventilation side of that equation with pulmonary blood flow presumed normal. Cardiac arrest and severe low-output states invert the problem: ventilation may be entirely normal (a rescuer or ventilator delivering full, unchanged tidal volumes) while pulmonary blood flow — and therefore CO2 delivery to the alveoli — collapses.
With little or no blood reaching the lungs, each delivered breath moves gas that contains little CO2, and the capnogram waveform shrinks dramatically in amplitude while retaining a roughly normal shape and rate (because the ventilator or bag continues cycling normally). This is the critical distinguishing feature from esophageal intubation (Stage 2): in low-output states the tube is correctly placed and a low-amplitude waveform with a recognizable shape persists, whereas esophageal placement typically shows the trace decay toward a true flat line with no residual periodicity.
During cardiac arrest with an advanced airway in place, capnography provides continuous, objective feedback that no other monitor can match:
• Compression quality: EtCO2 correlates directly with cardiac output generated by chest compressions. Deeper, faster, well-recoiled compressions with minimal interruption produce higher EtCO2; shallow or interrupted compressions produce low EtCO2. Rescuers can titrate compression technique in real time using the number itself, targeting EtCO2 above roughly 10–15 mmHg as a minimum and above 20 mmHg as a marker of good-quality CPR.
• Return of spontaneous circulation (ROSC): the abrupt native cardiac output generated by ROSC produces a sudden, often dramatic rise in EtCO2 — frequently a jump of 10 mmHg or more within a few beats — well before a pulse can be reliably palpated or a blood pressure cuff can cycle. This makes capnography one of the earliest and most reliable indicators that spontaneous circulation has returned during an ongoing resuscitation.
• Prognostic value: persistently low EtCO2 (below roughly 10 mmHg) despite twenty minutes of high-quality ACLS is a strong independent predictor of failure to achieve ROSC and is incorporated into some termination-of-resuscitation guidance, always alongside full clinical context rather than as a sole determinant.
• Confirmation of continued tube placement during resuscitation: because the airway is often placed emergently and the patient may be moved (transport, defibrillation positioning), a persistent capnographic waveform of any amplitude reassures that the advanced airway has not become dislodged, which a sudden true flat line would immediately reveal.
The 2020 AHA and ERC resuscitation guidelines both recommend continuous waveform capnography whenever an advanced airway is in place during CPR — both to confirm and continuously reconfirm correct tube placement, and to guide compression quality and help identify ROSC as early as possible.
The five patterns covered in this simulator represent the core waveform vocabulary every anesthesia and resuscitation provider should recognize instantly. A sixth pattern worth knowing by name — the "curare cleft" — appears as a small notch or dip in the middle of an otherwise normal Phase III plateau, produced by an early spontaneous inspiratory effort breaking through residual neuromuscular blockade; it signals returning respiratory muscle activity, useful when assessing readiness for extubation or adequacy of paralysis reversal.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| Normal (rectangular) | |||
| Absent / rapidly decaying | |||
| Shark-fin (sloped) | |||
| Elevated baseline | |||
| Rising plateau, clean baseline | |||
| Sudden low-amplitude | |||
| Curare cleft |