Capnography (CO2 Measurement, Mainstream, Sidestream)

Foundations and mechanisms

Physiological basis of expired carbon dioxide measurement

Capnography is the continuous measurement and graphical display of carbon dioxide concentration or partial pressure in respiratory gas. It is fundamentally a measurement of the interaction between CO₂ production, pulmonary blood flow, alveolar ventilation, ventilation–perfusion matching, and apparatus function. In anaesthetic practice, it is therefore both a respiratory and circulatory monitor.

Normal adult metabolic CO₂ production is approximately 200 ml min⁻¹ at rest, rising with fever, shivering, sepsis, malignant hyperthermia, tourniquet release and bicarbonate administration. CO₂ is transported predominantly as bicarbonate, with smaller fractions dissolved in plasma and bound to proteins. The dissolved component determines arterial carbon dioxide tension, normally PaCO₂ 4.7–6.0 kPa or 35–45 mmHg.

End-tidal carbon dioxide, PE′CO₂ or EtCO₂, approximates alveolar CO₂ at the end of expiration. In healthy anaesthetised adults it is typically 4.5–5.5 kPa or 34–42 mmHg. The normal arterial-to-end-tidal gradient is approximately 0.3–0.7 kPa or 2–5 mmHg, with PaCO₂ exceeding EtCO₂. The gradient widens when alveolar dead space increases, particularly in pulmonary embolism, hypovolaemia, low cardiac output, high airway pressures, emphysema, acute respiratory distress syndrome, and lateral or prone positioning.

Principles of CO₂ detection

Most clinical capnographs use infrared absorption spectroscopy. CO₂ is a heteronuclear linear molecule with strong infrared absorption around 4.26 μm. The Beer–Lambert law describes the relationship:

A = εcl, where A is absorbance, ε the molar absorptivity, c the concentration, and l the optical path length. In practice, emitted infrared radiation passes through a gas sample to a detector; reduction in transmitted radiation at the CO₂-specific wavelength is converted to partial pressure after calibration and compensation.

Important corrections include barometric pressure, water vapour, temperature, collision broadening by nitrous oxide and oxygen, and baseline drift. Water vapour is particularly relevant: saturated expired gas at 37°C contains water vapour pressure of 6.3 kPa or 47 mmHg, reducing the partial pressure available for other gases. Modern analysers report CO₂ as a dry or water-corrected value depending on device design.

Technology	Mechanism	Clinical relevance
Infrared absorption	CO2 absorbs IR light near 4.26 μm; signal obeys Beer–Lambert principles	Standard anaesthetic capnography; rapid, robust, gas-specific with compensation algorithms
Mass spectrometry	Ionises gas and separates by mass-to-charge ratio	Highly accurate multigas analysis; historically used in centralised systems, slower and expensive
Raman spectroscopy	Measures wavelength shift after laser scattering	Multigas capability; less common in routine theatres
Colorimetric detectors	pH-sensitive indicator changes colour with exhaled CO2	Qualitative confirmation of tracheal placement; unreliable in low pulmonary blood flow

Capnogram phases and waveform interpretation

The time capnogram is conventionally divided into phases that reflect sequential emptying of the respiratory tract. Understanding these phases is central to Primary FRCA viva and OSCE interpretation.

Phase	Physiological origin	Normal feature	Abnormal implications
Phase 0	Inspiratory baseline	CO2 approaches zero	Raised baseline suggests rebreathing, exhausted soda lime, incompetent valves, inadequate fresh gas flow
Phase I	Anatomical dead space gas	Near-zero CO2	Prolongation with increased apparatus dead space
Phase II	Mixing of dead space and alveolar gas	Rapid upstroke	Slurred upstroke with obstruction or sampling delay
Phase III	Alveolar plateau	Slight positive slope	Steep slope with V/Q inequality, bronchospasm, COPD, obesity, pregnancy
End-tidal point	Maximum expired alveolar CO2	Approximates PaCO2	Low in hyperventilation, low cardiac output, pulmonary embolism; high in hypoventilation or hypermetabolism

Mainstream and sidestream classifications

Capnographs are classified by the position of the analyser relative to the airway. In mainstream capnography, the infrared cuvette and detector are placed directly between the tracheal tube and breathing circuit. It provides a fast response, typically <100 ms, avoids sampling losses and is advantageous in small tidal volumes if dead space is acceptable. Disadvantages include added weight at the airway, risk of traction on the tracheal tube, heating requirement to prevent condensation, and added apparatus dead space, commonly 5–15 ml, clinically important in neonates and infants.

In sidestream capnography, gas is continuously aspirated through a sampling line to a remote analyser. Sampling flows are commonly 50–250 ml min⁻¹; low-flow neonatal systems may use 50 ml min⁻¹ or less. Sidestream devices allow convenient multigas analysis and use with facemasks, supraglottic airways and nasal cannulae, but introduce transit delay, waveform dispersion, vulnerability to obstruction by water or secretions, and dilution by leaks or high fresh gas flows. Response time depends on sampling flow, tubing length and analyser rise time; a delay of 1–4 s is typical.

Feature	Mainstream	Sidestream
Sensor location	At airway	Remote analyser via sampling tube
Response	Very rapid	Delayed and damped by transit and sampling dynamics
Dead space	Increased at airway	Minimal at airway
Water/secretions	Condensation affects optical window	Sampling line blockage common
Best suited to	Intubated patients, rapid waveform fidelity	Multigas monitoring, non-intubated sampling, flexible theatre use

Core numerical and exam-relevant thresholds

• Normal EtCO₂: 4.5–5.5 kPa, or 34–42 mmHg.
• Normal PaCO₂: 4.7–6.0 kPa, or 35–45 mmHg.
• Normal PaCO₂–EtCO₂ gradient: 0.3–0.7 kPa, or 2–5 mmHg.
• Infrared absorption peak of CO₂: approximately 4.26 μm.
• Water vapour pressure at 37°C: 6.3 kPa, or 47 mmHg.
• Sidestream sampling flow: commonly 50–250 ml min⁻¹.
• Mainstream airway dead space: typically 5–15 ml, important in paediatrics.

Contemporary standards from the Association of Anaesthetists, Difficult Airway Society and resuscitation guidelines regard continuous waveform capnography as mandatory for confirming and monitoring tracheal tube position during anaesthesia, sedation involving airway intervention, intensive care transfer and cardiac arrest when an advanced airway is present. Mechanistically, sustained exhaled CO₂ confirms ventilation of perfused alveoli; absence or abrupt loss of a waveform must be treated as oesophageal intubation, disconnection, obstruction, severe bronchospasm, or circulatory collapse until proven otherwise.

Clinical assessment and investigations

Clinical presentation: what capnography is expected to diagnose

Capnography is a continuous, breath-by-breath investigation of ventilation, perfusion and apparatus integrity, rather than a simple measure of arterial carbon dioxide. The displayed value, usually end-tidal carbon dioxide tension (P_ETCO₂), is interpreted with the waveform morphology, ventilatory settings, haemodynamics and, when required, arterial blood gas PaCO₂. Normal P_ETCO₂ in an anaesthetised adult is approximately 4.6–6.0 kPa or 35–45 mmHg. The normal PaCO₂–P_ETCO₂ gradient is 0.3–0.7 kPa (2–5 mmHg), but may exceed 1.3–2.0 kPa in major ventilation–perfusion mismatch, low cardiac output, pulmonary embolism or severe lung disease.

The principal clinical uses are confirmation of tracheal intubation, detection of circuit disconnection or obstruction, assessment of alveolar ventilation, monitoring during sedation, diagnosis of bronchospasm and rebreathing, and estimation of pulmonary blood flow during cardiopulmonary resuscitation. UK and international standards, including the Association of Anaesthetists and ASA Standards for Basic Anaesthetic Monitoring, require continuous waveform capnography for general anaesthesia and strongly recommend it for deep sedation, emergency airway management and transfer of ventilated patients.

Interpretation of the capnogram

A normal time capnogram has four phases: phase I anatomical dead space with near-zero CO₂; phase II rapid expiratory upstroke from mixed dead-space and alveolar gas; phase III alveolar plateau; and phase 0 inspiratory downstroke. The alpha angle between phases II and III is normally about 100–110°; it increases with expiratory flow limitation. The beta angle between phase III and inspiration is normally close to 90°; widening suggests rebreathing or valve malfunction.

Finding	Likely mechanism	Important differential diagnosis
Absent or near-flat waveform after intubation	No alveolar CO2 reaching sensor	Oesophageal intubation, complete circuit disconnection, apnoea, cardiac arrest with absent pulmonary blood flow, blocked tracheal tube, sampling-line obstruction
Sudden PETCO2 fall to zero	Loss of ventilation or sampling	Disconnection, extubation, ventilator failure, capnograph sampling pump failure, water trap full
Sudden fall with waveform persisting	Reduced pulmonary perfusion or increased dead space	Pulmonary embolism, severe hypotension, haemorrhage, cardiac arrest, high PEEP reducing venous return
Progressive rise in PETCO2	Increased CO2 production or reduced alveolar ventilation	Hypoventilation, malignant hyperthermia, sepsis, thyrotoxicosis, bicarbonate administration, tourniquet release, CO2 laparoscopy
Shark-fin upstroke and sloping plateau	Uneven alveolar emptying, increased airway resistance	Bronchospasm, COPD/asthma, kinked tube, mucus plug, inadequate expiratory time
Inspired CO2 above zero	Rebreathing	Exhausted soda lime, incompetent expiratory valve, insufficient fresh gas flow in Mapleson system, excessive apparatus dead space
Curare cleft	Spontaneous respiratory effort during mechanical ventilation	Inadequate neuromuscular blockade, light anaesthesia, ventilator dyssynchrony
Cardiogenic oscillations	Cardiac movement causing small gas shifts	Low respiratory rate, paediatric ventilation; usually benign but may mimic triggering

Investigations and confirmatory thresholds

Capnography is the reference bedside investigation for confirmation of tracheal tube placement: a sustained square waveform for at least 5–6 ventilated breaths is strong evidence of tracheal intubation. Transient CO₂ may occur after oesophageal intubation from swallowed gas or mask ventilation, so a disappearing waveform is unsafe. Conversely, false low or absent CO₂ can occur during cardiac arrest, severe bronchospasm, complete airway obstruction or profound low-flow states; therefore waveform interpretation must be integrated with chest movement, auscultation, airway pressure and oxygen saturation.

Arterial blood gas analysis is indicated when the PaCO₂–P_ETCO₂ gradient is clinically important: severe lung disease, one-lung ventilation, laparoscopic surgery, pregnancy with critical illness, head injury, cardiopulmonary bypass separation, and shock. During routine anaesthesia, ventilatory adjustment commonly targets P_ETCO₂ 4.5–5.5 kPa; neuroanaesthesia may target PaCO₂ 4.0–4.5 kPa when brief cerebral vasoconstriction is desired, avoiding profound hypocapnia because cerebral blood flow falls by approximately 2–4% per 1 mmHg PaCO₂ reduction within the autoregulatory range.

Critical-care, sedation and resuscitation interpretation

In procedural sedation, capnography detects hypoventilation and apnoea earlier than pulse oximetry, particularly when supplemental oxygen delays desaturation. Clinically significant respiratory depression is often defined in studies as apnoea >15–20 s, P_ETCO₂ >50 mmHg or an increase >10 mmHg from baseline, although thresholds are context-dependent and waveform loss is generally more actionable than an isolated number.

During CPR, P_ETCO₂ is a surrogate for pulmonary blood flow and chest compression quality. Values <10 mmHg after high-quality compressions are associated with very low likelihood of return of spontaneous circulation, whereas abrupt sustained rise to ≥35–40 mmHg suggests ROSC. Current resuscitation guidance cautions against using a single P_ETCO₂ cut-off alone to terminate resuscitation, but recommends using low values to prompt optimisation of compression depth, recoil, rate, ventilation and vasopressor timing.

Mainstream and sidestream sources of diagnostic error

Mainstream analysers give rapid response and minimal sampling delay but add apparatus dead space and weight at the airway, relevant in neonates and difficult airway fixation. Sidestream systems aspirate gas at approximately 50–200 mL min⁻¹; this may dilute small tidal volumes, create delay of about 1–4 s, and is vulnerable to leaks, water, secretions and long sampling tubing. High fresh gas flow, oxygen insufflation, non-invasive ventilation leaks and high respiratory rates can underestimate P_ETCO₂. Infrared analysers may be affected by collision broadening and spectral overlap, historically with nitrous oxide and volatile agents, although modern devices compensate algorithmically. Therefore the investigation is the capnogram plus clinical context, not the displayed number alone.

Management, pharmacology and procedures

Immediate clinical response to capnographic abnormalities

Capnography is a real-time safety monitor rather than a diagnostic test in isolation. Management is directed by the waveform first, then the numerical EtCO₂. A normal adult EtCO₂ is typically 4.7–6.0 kPa or 35–45 mmHg, with an arterial-to-end-tidal gradient of approximately 0.3–0.7 kPa in healthy anaesthetised patients, increasing with V/Q mismatch, low cardiac output, pulmonary embolism and lung disease.

Capnographic finding	Likely mechanism	Immediate management priorities
Absent waveform after intubation	Oesophageal intubation, circuit disconnection, complete airway obstruction, cardiac arrest, sampling failure	Assume misplaced tube until proven otherwise; ventilate with 100% O2, inspect chest movement, auscultate, check circuit and capnograph, perform direct/video laryngoscopy or fibreoptic confirmation; remove tube if doubt persists
Sudden fall to zero	Disconnection, extubation, ventilator failure, sampling line dislodgement	Check patient first, then airway, breathing system, ventilator, sampling line and water trap
Progressive rise in EtCO2	Hypoventilation, increased CO2 production, rebreathing, exhausted soda lime, malignant hyperthermia	Increase minute ventilation; check fresh gas flow, valves and absorber; assess temperature, rigidity, tachycardia, acidosis and potassium
Progressive fall in EtCO2	Hyperventilation, falling cardiac output, pulmonary embolism, severe hypotension, hypothermia	Assess circulation; treat hypotension, haemorrhage, embolic event or cardiac arrest; avoid interpreting EtCO2 as PaCO2 when dead space is increased
Shark-fin expiratory upstroke	Expiratory flow limitation: bronchospasm, COPD, kinked tube, secretions	Deepen anaesthesia, give bronchodilator, suction airway, check tube/circuit obstruction and ventilatory pressures

Procedural application and equipment management

Continuous waveform capnography is mandatory for tracheal intubation and maintenance of general anaesthesia, and is recommended during deep sedation, interhospital transfer and CPR. UK guidance following NAP4 emphasised that failure to use or interpret capnography contributed substantially to airway-related morbidity and mortality. During CPR, a persistent EtCO₂ <10 mmHg after high-quality compressions suggests poor pulmonary blood flow and should prompt optimisation of compression depth, rate, recoil and ventilation; an abrupt sustained rise, often to >35–40 mmHg, strongly suggests ROSC. It should not be used as the sole criterion to terminate resuscitation.

Mainstream analysers should be placed between the airway device and breathing system, minimising dead space and avoiding excessive weight on the tracheal tube. Sidestream systems aspirate gas at commonly 50–250 ml min^-1; this is clinically important in neonates, small infants and low-flow anaesthesia. The sampling line should be connected as close to the airway as practicable, with a water trap/filter, and should be checked for obstruction, leaks and excessive transit delay. Calibration is usually automatic against a reference cell or room air depending on device design; zeroing and leak checks are required after sensor replacement, contamination or unexpected readings.

Pharmacological and therapeutic responses guided by capnography

Capnography frequently provides the earliest warning of drug-induced ventilatory depression. Opioids, benzodiazepines, propofol and volatile agents reduce ventilatory drive and increase PaCO₂; management is airway support and dose reduction before antagonist therapy. If reversal is required, naloxone is titrated intravenously in 20–40 microgram aliquots in opioid-dependent or postoperative patients, or 100–200 microgram increments in severe respiratory depression; duration is 30–90 minutes, shorter than many opioids, so recurrence requires observation or infusion. Flumazenil 200 microgram IV over 15 seconds, repeated every 60 seconds to 1 mg, may reverse benzodiazepine sedation but risks seizures in dependence or mixed overdose.

A rapidly rising EtCO₂ despite increased ventilation is a classic early sign of malignant hyperthermia. Management is immediate discontinuation of volatile anaesthetics and suxamethonium, 100% oxygen at high fresh gas flow, active cooling, correction of hyperkalaemia/acidosis and dantrolene 2.5 mg kg^-1 IV repeated until control, often requiring cumulative doses 10 mg kg^-1 or more. Post-crisis dantrolene is commonly continued 1 mg kg^-1 every 4–6 hours or infusion for 24–48 hours, with ICU monitoring for recrudescence.

Bronchospasm causing a sloping phase II/III waveform is treated by deepening anaesthesia, removing triggers and administering inhaled salbutamol 4–10 puffs via spacer into the circuit or nebulised 2.5–5 mg; severe cases may require adrenaline 10–100 microgram IV titrated cautiously, or 1 mg IM in anaphylaxis with circulatory compromise, alongside fluids and antihistamine/steroid adjuncts.

Complications, artefact and follow-up

Capnography-related complications are usually interpretative or equipment-related. Sidestream sampling may dilute readings through leaks, entrain room air during high oxygen flows, or underestimate EtCO₂ if sampling is distal from the airway. Water, secretions and blood obstruct sampling lines and generate falsely low or absent traces. Mainstream sensors add apparatus dead space, relevant in small children, and may be affected by condensation or contamination. Nitrous oxide, oxygen and volatile agents can cause spectral interference, though modern infrared analysers compensate using multi-wavelength algorithms.

Following any critical capnographic event, documentation should include waveform description, EtCO₂ value and trend, airway findings, interventions and response. Recurrent unexplained hypercapnia warrants review of ventilator settings, absorber integrity, circle valves and metabolic causes; recurrent large PaCO₂–EtCO₂ gradients should prompt assessment for dead-space pathology, including pulmonary embolism, low cardiac output and severe obstructive lung disease.

Exam controversies and advanced synthesis

Guideline position: capnography is no longer “optional monitoring”

For examination purposes, waveform capnography should be regarded as a standard of care whenever ventilation is intentionally controlled or airway patency is at risk. The Association of Anaesthetists’ monitoring standards require continuous capnography for general anaesthesia from induction until recovery of airway reflexes, for patients with supraglottic devices, tracheal tubes and during transfer of anaesthetised or ventilated patients. ASA standards similarly mandate continual end-tidal CO₂ analysis during general anaesthesia and require exhaled CO₂ monitoring during moderate or deep sedation unless precluded by patient, procedure or equipment limitations.

The UK NAP4 audit is frequently cited in vivas: failure to use capnography, particularly in ICU and emergency departments, contributed to delayed recognition of oesophageal intubation, airway obstruction and displaced tracheostomies. The key exam message is that a stethoscope, chest movement, reservoir bag movement, fogging or pulse oximetry are inferior because they are delayed, indirect or non-specific. Persistent waveform capnography is the most reliable clinical confirmation of tracheal ventilation.

Clinical setting	Expected exam answer	Common pitfall
Tracheal intubation	Sustained square waveform for at least 6 breaths confirms tracheal ventilation	Transient CO2 after oesophageal intubation from gastric insufflation
Procedural sedation	Capnography detects hypoventilation before SpO2 falls, especially with supplemental O2	Assuming normal SpO2 excludes ventilatory depression
ICU/transfer	Continuous waveform capnography for all mechanically ventilated patients	Relying on ventilator pressure alarms after disconnection or displaced airway
CPR	Assesses tube position, compression quality and ROSC	Using a single EtCO2 value alone to terminate resuscitation

Evidence and controversies

Randomised trials and meta-analyses in procedural sedation consistently show that capnography reduces episodes of hypoxaemia and detects respiratory depression earlier than pulse oximetry; however, evidence for mortality reduction is limited because catastrophic airway events are rare. This distinction is viva-relevant: the absence of mortality trials does not imply absence of benefit, since capnography is a high face-validity safety monitor with strong mechanistic plausibility and audit evidence.

In cardiac arrest, EtCO₂ approximates pulmonary blood flow when ventilation and metabolism are constant. Values <1.3 kPa or <10 mmHg after approximately 20 minutes of high-quality CPR are associated with poor outcome, whereas an abrupt rise to approximately 4.0–5.3 kPa or 30–40 mmHg suggests ROSC. Current resuscitation guidance cautions that EtCO₂ should not be used in isolation to stop CPR because it is confounded by adrenaline, bicarbonate, ventilation rate, pulmonary embolism, airway leak, low temperature and compression quality.

Viva-level interpretation: integrate waveform, number and context

The normal adult EtCO₂ is approximately 4.6–6.0 kPa or 35–45 mmHg. The normal PaCO₂–EtCO₂ gradient is about 0.3–0.7 kPa or 2–5 mmHg, reflecting alveolar dead space. The gradient widens with increased V/Q mismatch, low cardiac output, pulmonary embolism, emphysema, severe hypovolaemia and during CPR. Thus EtCO₂ is a surrogate for PaCO₂, not a substitute in unstable patients.

• High EtCO₂ with rising baseline: rebreathing from exhausted soda lime, incompetent valves, inadequate fresh gas flow, excessive apparatus dead space or faulty circle system.
• High EtCO₂ with normal baseline: hypoventilation, increased CO₂ production, sepsis, fever, shivering, thyrotoxicosis, tourniquet release, bicarbonate administration or malignant hyperthermia.
• Low EtCO₂: hyperventilation, low cardiac output, pulmonary embolism, severe hypotension, hypothermia, disconnection, airway leak or sampling failure.
• Sloping phase III: obstructive disease, bronchospasm, kinked tube, endobronchial intubation or heterogeneous alveolar emptying.
• Absent waveform: oesophageal intubation, apnoea, complete obstruction, disconnection, cardiac arrest with negligible pulmonary blood flow, or monitor/sampling malfunction.

Mainstream versus sidestream: advanced pitfalls

Mainstream analysers have rapid response and avoid sampling-line delay, but add weight and dead space at the airway, increasing traction and potentially affecting neonates or low tidal volume ventilation. Sidestream systems are versatile and compatible with non-intubated sampling, but aspirate gas at typically 50–250 mL min⁻¹. This may be significant in neonates or during low-flow ventilation, and causes delay, dispersion and dilution, especially with long sampling lines, high respiratory rates, water contamination or leaks.

Infrared analysers exploit CO₂ absorption near 4.26 μm and obey Beer–Lambert principles, but measurement is affected by pressure, temperature, water vapour, nitrous oxide and collision broadening. Modern monitors compensate electronically, yet exam answers should mention calibration/zeroing, water traps, blocked filters, cracked sampling lines and sample dilution by oxygen insufflation. A falsely low EtCO₂ in a sedated patient receiving high-flow nasal oxygen may reflect sampling dilution rather than adequate ventilation.

Classic examination traps

1. “Capnography confirms tube position.” More precisely, sustained waveform confirms ventilation of CO₂-containing lung; it does not exclude endobronchial intubation.
2. “Normal EtCO₂ means normal PaCO₂.” False in major V/Q mismatch or low cardiac output; arterial blood gas measurement may be required.
3. “Pulse oximetry is enough during sedation.” False: oxygenation and ventilation are distinct, and supplemental oxygen delays desaturation.
4. “Absent CO₂ always means oesophageal intubation.” Consider cardiac arrest, profound hypotension, complete obstruction, disconnection or analyser failure; nevertheless, after intubation it must be treated as oesophageal until proven otherwise.