Continuous Waveform Capnography in Patient Monitoring

Continuous waveform capnography is real-time monitoring of the carbon dioxide concentration in a patient’s exhaled breath, displayed as a rolling waveform on a screen. Unlike a simple numeric readout of end-tidal CO₂, the continuous waveform plots CO₂ levels against time with every breath, creating a characteristic shape that tells clinicians far more than a single number can. Because CO₂ production is tied to metabolism, circulation, and ventilation simultaneously, that waveform has become one of the most versatile monitoring tools in emergency medicine, anesthesia, and critical care.

What the Waveform Actually Shows

A normal capnography waveform looks like a rounded rectangle that repeats with each breath. It has distinct phases, each reflecting a different stage of exhalation. Phase I is a flat baseline near zero, representing the air coming from the upper airways and trachea where no gas exchange happens. Phase II is a sharp upswing as air from deeper in the lungs, where CO₂ has been exchanged, begins mixing with that dead-space gas. Phase III is the “alveolar plateau,” a relatively flat stretch at the top of the waveform representing CO₂-rich air from the alveoli. The highest point at the end of Phase III is the end-tidal CO₂ value. The waveform then drops steeply back to baseline as the patient inhales fresh gas.

The angles between these phases carry clinical meaning. The angle where Phase II meets Phase III (the alpha angle) reflects how evenly the lungs are emptying. A steeper or prolonged upslope suggests airway obstruction, as different lung regions empty at different rates. In some patients, a Phase IV appears as a small uptick right at the end of Phase III, seen in conditions like pregnancy or obesity where closing lung volumes shift.

¹ The shape of that waveform, not just the peak number, is what makes continuous capnography so much more informative than a colorimetric detector or a single spot-check reading.

Confirming an Airway Is in the Right Place

The most established use of continuous waveform capnography is confirming that a breathing tube sits in the trachea rather than the esophagus. A tube accidentally placed in the esophagus delivers no ventilation to the lungs, and if unrecognized, the consequences are catastrophic. Traditional methods like listening for breath sounds or watching for chest rise can be unreliable, especially in noisy or chaotic settings. Capnography cuts through the ambiguity: a sustained, repeating waveform with the classic four-phase shape means gas is moving in and out of functioning alveoli. No waveform, or a rapidly flattening one, means the tube is in the wrong place.

A cadaveric study tested this rigorously, delivering 195 tracheal ventilations and 48 extratracheal ventilations (esophageal and hypopharyngeal combined). Every tracheal ventilation produced a characteristic alveolar waveform. Every extratracheal ventilation produced a flat, non-alveolar tracing with CO₂ readings of zero. The result was 100% sensitivity and 100% specificity for distinguishing tracheal from extratracheal tube placement using waveform analysis.

A systematic review of prehospital intubation confirmation methods similarly found that capnography showed high sensitivity and specificity for verifying tube placement in out-of-hospital settings.

Capnography does not just confirm initial placement. It also catches problems that develop after the tube is secured. An animal study comparing capnography to pulse oximetry and vital signs found that when a tube slipped out of the trachea into the upper throat, capnography detected the problem within about seven seconds by showing a flattened waveform, far faster than pulse oximetry or vital sign changes.

⁴ That speed advantage matters most when patients have been pre-oxygenated, because oxygen saturation can remain normal for minutes even after ventilation has stopped, masking a dislodged tube.

Guiding CPR and Predicting Who May Survive

During cardiac arrest, continuous capnography serves a completely different purpose. When the heart stops pumping, CO₂ still gets produced by tissues but cannot reach the lungs effectively. Chest compressions act as a crude pump, pushing some blood through the lungs, and the amount of CO₂ that makes it to the exhaled breath reflects how well that chest-compression-driven circulation is working. The end-tidal CO₂ reading during CPR therefore becomes a real-time proxy for cardiac output.

Research has quantified this relationship. For every additional centimeter of chest compression depth, the end-tidal CO₂ rose by about 1.4 mmHg. Higher ventilation rates, on the other hand, diluted the CO₂ and lowered the reading by about 3 mmHg for every additional 10 breaths per minute.

⁵ A study in pediatric cardiac arrest found that average end-tidal CO₂ values of 20 mmHg or higher were associated with higher blood pressures during CPR and shorter pre-arrest durations, though compression rate and depth did not differ between the groups, suggesting the reading reflects underlying patient physiology as much as technique.

⁶

Perhaps the most watched moment during resuscitation is when end-tidal CO₂ suddenly jumps. A sharp rise often signals the return of a heartbeat, because the restored circulation sweeps a bolus of accumulated CO₂ from the tissues to the lungs all at once. A study of out-of-hospital cardiac arrest patients with pulseless electrical activity found that an increase in end-tidal CO₂ greater than 20 mmHg was highly specific for return of spontaneous circulation, leading the authors to recommend that resuscitation continue in patients with rising values unless overwhelming clinical evidence says otherwise.

⁷

Conversely, persistently low end-tidal CO₂ during prolonged resuscitation suggests poor perfusion and a grim prognosis. Tracking trajectories over time adds prognostic power. An emergency department study found that patients whose end-tidal CO₂ trended along a 30 mmHg trajectory were roughly five times more likely to regain a pulse than those tracking at 10 mmHg, and those at 50 mmHg were about eight times more likely.

⁸ Combining end-tidal CO₂ data with clinical scoring improved the ability to predict outcomes after out-of-hospital arrest beyond what either measure achieved alone.

⁹

Procedural Sedation and the General Care Floor

Outside the operating room and the cardiac arrest bay, continuous capnography has found a growing role in monitoring patients who are sedated for procedures like fracture reduction, abscess drainage, or endoscopy. The premise is straightforward: sedation drugs suppress breathing, and a drop in respiratory effort will show up on the capnography waveform before oxygen saturation falls on a pulse oximeter. Early detection should, in theory, allow clinicians to intervene sooner and prevent dangerous oxygen drops.

A prospective study in children undergoing procedural sedation found that capnography did detect respiratory depression earlier than pulse oximetry alone.

¹⁰ The evidence in adults is less clear-cut. A Cochrane systematic review of emergency department procedural sedation found no convincing evidence that adding capnography to standard monitoring reduced rates of oxygen desaturation, low blood pressure, or vomiting. It even noted that capnography may increase the rate of airway interventions in adults, possibly because clinicians respond to waveform changes that would have resolved on their own.

¹¹

On general hospital floors, a different population stands to benefit: patients receiving opioid painkillers after surgery. Opioid-induced respiratory depression can happen hours after a dose, often when the patient is asleep and unobserved. A cost-modeling study based on a large trial estimated that continuously monitoring high-risk patients with both capnography and pulse oximetry could reduce annual hospital costs by over half a million dollars and cut cumulative extra hospital days by more than a hundred, primarily by catching respiratory depression before it escalates to a code blue or ICU transfer.

¹²

Diagnostic Clues Hidden in the Waveform

Beyond confirming tube placement and monitoring ventilation, the shape and behavior of the capnography waveform can point toward specific diagnoses. Features of the CO₂ waveform have been shown to correlate with ventilation-perfusion mismatch, dead-space volume, breathing patterns, and small airway obstruction.

¹³

In obstructive lung diseases like asthma and COPD, the alveolar plateau (Phase III) loses its flat shape and slopes upward, sometimes described as a “shark fin” pattern. This happens because narrowed airways cause different lung regions to empty at different speeds, so the CO₂ concentration keeps climbing throughout exhalation instead of leveling off. The more severe the obstruction, the steeper the slope. Researchers have explored using machine learning on high-resolution capnography data to diagnose COPD, suggesting these waveform features carry enough information to support automated screening.

Pulmonary embolism offers another example. When a blood clot blocks flow to part of the lung, that region still receives air but has no blood flow to exchange CO₂ with, creating dead space. The result is a widened gap between arterial CO₂ (measured by blood draw) and end-tidal CO₂ (measured by capnography). A review of clinical studies concluded that this gap has good evidence supporting its use to help exclude pulmonary embolism when pretest probability is low, though it lacks the specificity to confirm the diagnosis on its own.

¹⁴ When thrombolytic therapy succeeds in breaking up a pulmonary embolism, capnography can track the response in real time: as blood flow returns to previously blocked lung segments, the arterial-to-end-tidal CO₂ gap narrows, offering a noninvasive bedside gauge of treatment effectiveness.

¹⁵

Fluid Responsiveness in Shock

One of the more recent applications is using changes in end-tidal CO₂ to predict whether a patient in shock will benefit from intravenous fluids. The logic is simple: if giving fluid increases cardiac output, more CO₂ gets delivered to the lungs and the end-tidal value rises. A passive leg raise, which temporarily shifts blood from the legs to the heart, can be used as a reversible “fluid challenge” while watching the capnography reading.

A prospective study in septic shock patients found that a change in end-tidal CO₂ of 5% or more during a passive leg raise predicted fluid responsiveness with a specificity of over 99% and sensitivity of about 76%.

¹⁶ A study in cardiogenic shock found that the change in end-tidal CO₂ outperformed other common markers like blood pressure changes and inferior vena cava measurements in predicting who would respond to fluids.

¹⁷ A systematic review and meta-analysis confirmed that end-tidal CO₂ changes during passive leg raising showed moderate predictive ability in mechanically ventilated patients, though it acknowledged that the number of studies was still limited and heterogeneous.

¹⁸

Technology Differences That Matter Clinically

Three main sensor designs exist for continuous capnography, and each has trade-offs that affect where it works well and where it falls short. Mainstream sensors sit directly in the breathing circuit between the tube and the ventilator. They measure CO₂ right at the source, producing fast and accurate waveforms with minimal delay. The downside is a heavier connector at the airway, which can torque on a small patient’s tube, and the sensor must be heated to prevent condensation, adding bulk and fragility.

Sidestream sensors pull a small sample of gas through thin tubing to an analyzer housed in the monitor. This allows a lighter connection at the patient and makes it possible to monitor non-intubated patients using a nasal cannula adapter. The penalty is a slight time delay and the potential for the sampling line to become blocked by moisture or secretions. Conventional sidestream and mainstream systems both rely on infrared absorption to detect CO₂, and the technical limitations of that approach historically kept capnography tethered to operating rooms and ICUs.

A third approach, developed more recently, uses molecular correlation spectroscopy to measure CO₂. This technology generates a highly efficient and selective emission spectrum matched precisely to CO₂ absorption wavelengths, requiring much smaller sample volumes and lower flow rates than conventional sidestream systems.

¹⁹ The result is a smaller, lighter device that handles the low flow rates needed for neonates and the portability demanded by prehospital environments. Compact and rugged capnography units with extended battery life are now considered essential equipment for prehospital advanced airway management.

²⁰

Where Capnography Misleads

For all its strengths, continuous capnography has real blind spots. End-tidal CO₂ is not the same as arterial CO₂, and the gap between the two widens unpredictably in patients with significant lung disease. In ventilated neonates, this is a particular concern. The small tidal volumes and rapid breathing rates of newborns can overwhelm sensor response times, and apparatus dead space from the sensor itself becomes proportionally much larger relative to the tiny volumes being measured. If technical specifications are not carefully matched to the patient’s size, the readings can be misleading.

²¹ A study of sidestream capnography in ventilated infants found that the correlation between end-tidal and arterial CO₂ was reasonable in babies with less severe lung disease but degraded substantially in those with more severe disease and greater ventilation-perfusion mismatch.

²²

The tube-placement confirmation role, while remarkably reliable, is not immune to false positives. If a patient has recently consumed a carbonated drink and the tube ends up in the esophagus, CO₂ from the beverage in the stomach can produce a waveform that initially mimics a tracheal placement. An experimental study found that even small amounts of carbonated beverage in the gastric sac caused a false-positive result on a colorimetric capnometer for up to ten breaths.

²³ An animal study confirmed this, observing CO₂ waveforms during esophageal ventilation after carbonated beverages were placed in the stomach, with end-expired CO₂ levels reaching as high as 5.3% in some cases.

²⁴ This is why experienced clinicians look for a sustained, repeating waveform over multiple breaths rather than trusting the first few cycles alone. A gastric CO₂ source will wash out quickly, while a tracheal waveform persists.

The Cost Question

Whether routine capnography monitoring makes financial sense depends heavily on the clinical context. During gastrointestinal endoscopy under deep sedation, a cost-benefit model estimated that capnography reduced the cost per procedure by about $85, because catching adverse events early enough to prevent escalation saved more than the monitoring equipment cost.

²⁵

In emergency department procedural sedation, the picture is starkly different. A systematic review and cost-effectiveness analysis estimated that routine capnography monitoring in ED sedations would cost roughly $2.8 million per prevented catastrophic event, translating to about $114,000 per quality-adjusted life-year.

²⁶ The difference comes down to baseline risk: endoscopy sedation is deeper, longer, and involves patients who are often older and sicker, so the monitoring prevents more harm per dollar spent. ED sedation for a young person getting a dislocated shoulder reduced already carries a low catastrophic event rate, so the monitoring cost spreads over very few prevented events.

These numbers do not mean capnography is useless in the ED. They mean that a blanket policy of monitoring every sedation equally may not be the smartest allocation. Targeting monitoring toward higher-risk patients and deeper sedation levels is where the value concentrates, a point the endoscopy modeling data reinforces: cost savings dropped from $85 per procedure under deep sedation to $35 under moderate sedation. Because CO₂ monitoring reflects metabolism, circulation, and ventilation all at once, the clinical settings where those three variables are most unstable are where the waveform earns its keep.

²⁷