Read the physiology.
Understand the limit.

Cardiopulmonary exercise testing combines progressively increasing exercise with measurement of oxygen uptake (V̇O₂), carbon dioxide output (V̇CO₂), and ventilation (V̇E), alongside ECG, blood pressure, symptoms, and oxygen saturation. A treadmill stress ECG asks a narrower question. A six-minute walk measures performance in a different setting. Neither supplies the same breath-by-breath physiology.

Use CPET to quantify exercise capacity, investigate unexplained exertional symptoms, characterize a known disease’s functional consequences, or guide selected treatment and rehabilitation decisions. Before ordering it, finish this sentence: “If the pattern is ___, the next decision will be ___.” A test ordered only to obtain a fitness number may not resolve the patient’s actual problem.

Separate three conclusions

• Observation: what was measured, under which conditions?
• Physiological interpretation: which response appears abnormal or limiting?
• Clinical diagnosis: what additional evidence identifies its cause?

“Reduced capacity with inefficient ventilation” is an interpretation. “Pulmonary arterial hypertension” is a diagnosis that CPET alone cannot establish. Similarly, preserved peak capacity does not exclude every exercise-related disorder, particularly when the patient’s usual symptoms were not reproduced.

CaO₂ and CvO₂ denote arterial and mixed venous oxygen content. Exercise demands coordinated ventilation, gas transfer, pulmonary perfusion, cardiac output, blood oxygen carriage, and peripheral extraction. Failure at several points can produce low V̇O₂. The measurement is integrated; it does not identify a single failing organ by itself.

Oxygen pulse is V̇O₂ divided by heart rate. Algebraically, it equals stroke volume multiplied by arteriovenous oxygen-content difference. It is therefore not a direct stroke-volume measurement. Hemoglobin, saturation, extraction, rhythm, and averaging all affect it. A rising heart rate with a flattening oxygen pulse is a reason to examine the other curves, not proof of ischemia.

Remember the units. V̇O₂ in mL/min divided by beats/min gives mL/beat. If cardiac output is in L/min and oxygen content is in mL/dL, a unit conversion is required in the Fick calculation. Avoid apparent physiological abnormalities created by mixing L and mL.

Before the patient starts

Record the indication, usual limiting activity, recent clinical stability, exercise familiarity, medications and their timing, oxygen use, and relevant resting tests. Decide whether the question is response on usual treatment or response under a specifically prescribed medication protocol. Do not routinely withhold beta blockers to make a predicted heart-rate target easier to reach.

Confirm gas and flow calibration, analyzer timing, mask or mouthpiece seal, ergometer function, ECG quality, and oximeter signal. Document oxygen delivery and equipment compatibility when supplemental oxygen is used. Gas fractions, leaks, and calibration errors can change calculated uptake.

Protocol

A cycle ramp gives a defined external work rate; treadmill testing often achieves a higher peak uptake and requires different reference equations. Select the increment to make the incremental phase approximately 8–12 minutes when feasible, after baseline and unloaded exercise. Include monitored recovery. Record actual duration, cadence, assistance, and reason for stopping. A fixed ramp for every patient risks either premature leg fatigue or an unnecessarily long test.

Safety is a separate clinical responsibility

Use a staffed laboratory with emergency equipment, resuscitation capability, and a locally approved screening and termination protocol. Acute coronary syndromes, decompensated heart failure, unstable significant arrhythmias, acute pulmonary embolism, acute myocarditis, and symptomatic severe aortic stenosis are examples requiring deferral or specialist reassessment—not an exhaustive contraindication list.

Chest pain with concerning ischemic changes, hemodynamic deterioration, significant arrhythmia, presyncope, severe oxygen desaturation, technical failure affecting safe monitoring, or the patient’s request can require stopping. Apply the laboratory’s indication-specific criteria. Reaching a predicted heart rate or an RER threshold is not, by itself, a reason to terminate a symptom-limited test.

Inspect the raw trends before reading the automated summary. Look for abrupt discontinuities, impossible resting values, sudden loss of ventilation, implausible gas fractions, a mismatched exercise-end marker, and isolated saturation dips. Correlate anomalies with technician notes and symptoms.

RER is V̇CO₂/V̇O₂. A peak near or above 1.10 commonly supports strong effort, but is neither necessary in every patient nor sufficient on its own. Hyperventilation can increase it, while severe ventilatory limitation or an early safety stop can prevent it rising. Combine RER with symptoms, perceived exertion, workload, physiological responses, and the stopping reason. Write “terminated for knee pain; maximal capacity not established” when that is what happened.

State the averaging interval used for peak values and inspect whether smoothing hides transitions or creates apparent plateaus. A single breath is not a reliable peak. Identify which conclusions survive a submaximal study; do not convert all remaining measurements into a normal result.

The 2026 ERS reference-values standard highlights limitations in available normative data. Select a reference appropriate to age, body characteristics, modality, and population; report its name. Percent predicted is not a universal scale interchangeable between laboratories.

Report peak V̇O₂ in L/min, mL/kg/min, and percent predicted, with the reference equation and averaging period. “Peak” is appropriate unless a true maximal plateau has been demonstrated. Where validated reference limits or z-scores exist, use them; 80% or 85% predicted are common orientation points, not universal diagnostic boundaries.

Weight indexing can make capacity look disproportionately low in obesity. Compare absolute uptake, work achieved, and suitable predicted values before attributing the entire deficit to cardiac disease. Conversely, a high percent predicted does not explain a substantial decline in an unusually fit person. Weight loss can increase mL/kg/min without increasing absolute uptake.

Uptake versus work

During appropriately performed cycle exercise, the V̇O₂–work-rate relationship is often around 10 mL/min/W. Examine its slope and shape after the early transition, accounting for the delay in oxygen uptake during a ramp. A persistently shallow relationship or late flattening invites review of oxygen delivery, symptoms, ECG, and measurement validity. It is not a standalone coronary test.

Report a functional limitation separately from an etiologic claim. The patient may have reduced performance with no single dominant cardiopulmonary pattern. Anemia, peripheral limitation, pain, obesity, and low conditioning remain clinical possibilities—not diagnoses assigned by exclusion from one curve.

The first ventilatory or gas-exchange threshold (VT1/GET, often labeled AT) reflects a change in the relationship between oxygen uptake, carbon dioxide output, and ventilation as exercise intensity rises. It is not a moment when aerobic metabolism stops or the muscles suddenly “run out of oxygen.” Gas exchange estimates a physiological transition; it does not directly measure blood lactate.

• Look for a disproportionate rise in V̇CO₂ relative to V̇O₂ on the V-slope plot.
• Seek corroboration: V̇E/V̇O₂ and end-tidal O₂ rise while V̇E/V̇CO₂ has not yet risen and end-tidal CO₂ is relatively maintained.
• Separate the later respiratory compensation point, where ventilation increases relative to CO₂ output and end-tidal CO₂ falls.

Report the threshold’s V̇O₂ and associated work rate, plus confidence in identification. Expressing it as a fraction of predicted peak capacity avoids making an early threshold appear reassuring simply because achieved peak uptake was very low. If irregular breathing, early termination, oscillation, or noise prevents identification, say so rather than forcing a number.

A reliably identified threshold can inform training intensity in a stable patient. Transfer workload or heart rate to the training setting cautiously: modality, medications, heat, symptoms, and recovery still matter.

Describe the starting rhythm, its evolution during exercise, peak rate, rate relative to metabolic demand, and recovery. A failure to reach 85% of age-predicted maximum is not an automatic diagnosis of chronotropic incompetence. Prediction equations have wide individual error; effort, beta blockers, other rate-limiting drugs, and early stopping matter.

One commonly used calculation is achieved heart-rate reserve: (peak HR − resting HR)/(predicted maximum HR − resting HR). Its interpretation requires a stated prediction equation and a definition validated for the population and medication context. Do not use a universal threshold for patients in AF, on rate-limiting drugs, and in sinus rhythm as though these were equivalent tests.

In AF, irregular cycle lengths and changing rate can distort short averages and oxygen-pulse curves. Correlate the tracing with symptoms and uptake rather than diagnosing a stroke-volume plateau from a noisy ratio. In a paced patient, inspect the device’s upper-rate behavior and rate-response settings with the device team; CPET can expose a rate ceiling but cannot prescribe programming by itself.

Systolic pressure ordinarily rises with exercise. An unexpected fall, a markedly excessive response, ischemic ECG changes, or clinically important arrhythmia changes the interpretation and may change test safety. A normal exercise ECG does not exclude coronary disease. Record recovery conditions before interpreting heart-rate recovery; active and passive recovery are not interchangeable.

The V̇E–V̇CO₂ slope describes the rise in ventilation relative to CO₂ output. State whether it was fitted through the full exercise period or only before respiratory compensation; the methods can give different numbers. The slope, the nadir V̇E/V̇CO₂ ratio, and a peak ratio are different measurements.

High ventilation for a given CO₂ output can reflect increased physiological dead space, a lower arterial CO₂ set point, or both. Heart failure, pulmonary vascular disease, lung disease, and hyperventilation can overlap here. A slope above roughly 35 is frequently a concern in adult clinical practice, but age, reference data, method, and disease context determine its meaning.

Follow end-tidal CO₂ (P_ETCO₂) from rest through the threshold and peak. Low values with an absent early rise, particularly alongside high ventilatory inefficiency and impaired oxygen delivery, increase concern about pulmonary vascular physiology. Low values also occur with hyperventilation. End-tidal CO₂ is not an interchangeable substitute for PaCO₂ when ventilation–perfusion matching is abnormal.

Do not calculate true dead-space fraction from end-tidal CO₂ alone. An Enghoff estimate uses arterial CO₂ and mixed expired CO₂ and reflects more than anatomical dead space. Specify sampling and calculation methods. Exercise oscillatory ventilation is another distinct finding: describe persistent cyclic fluctuation, verify it is not artifact, and apply a published definition before assigning prognostic meaning.

Use the resting pulmonary tests correctly

Interpret FEV₁/FVC against the appropriate lower limit of normal. Low FVC alone does not establish restriction; a reduced total lung capacity (TLC) is needed. An elevated residual-volume/TLC relationship can support air trapping. DLCO measures gas transfer and should be interpreted with hemoglobin adjustment and test quality in mind. A low value can accompany emphysema, interstitial disease, pulmonary vascular disease, or anemia. It does not identify which one.

DLCO, TLC, and spirometry are separate measurements, not outputs calculated from CPET. Normal spirometry does not establish normal lung volumes or normal gas transfer. Interpret DLCO alongside alveolar volume and clinical information; a seemingly preserved KCO does not automatically “correct” a low DLCO.

Breathing reserve

Calculate breathing reserve as 100 × (MVV − peak V̇E)/MVV. State whether maximal voluntary ventilation was measured or estimated, commonly from FEV₁ using a factor around 35–40. Values near or below 15–20% may suggest little remaining ventilatory capacity, but the estimate is imprecise and healthy highly trained people can use much of their reserve.

Dynamic mechanics can matter even when this reserve appears preserved. Examine tidal-volume expansion, respiratory frequency, tidal flow–volume relationships, and serial inspiratory-capacity maneuvers when available. A fall in inspiratory capacity during exercise can support dynamic hyperinflation if TLC is stable and maneuvers are reliable. Approaching a low inspiratory reserve can produce disproportionate dyspnea; the patient may stop before exhausting an estimated MVV.

Oxygen saturation and airway symptoms

Report baseline, nadir, and peak SpO₂, oxygen supplementation, and signal quality. Reproducible desaturation deserves attention, but movement, poor perfusion, and device limitations can mislead. Arterial sampling may be necessary when the gas-exchange question requires precision. Interpret any alveolar–arterial oxygen gradient with age, inspired oxygen, workload, and sampling conditions.

A standard incremental CPET does not reliably exclude exercise-induced bronchoconstriction. When clinically indicated, use an appropriate exercise or bronchoprovocation protocol with serial post-challenge FEV₁. The ATS guideline uses a fall of at least 10% from baseline; symptoms alone are insufficient. Inspiratory noise with abrupt exercise symptoms raises a different question, including exercise-induced laryngeal obstruction and possible continuous laryngoscopy during exercise.

Dysfunctional breathing is not synonymous with anxiety. It can coexist with disease, and CPET lacks a single universally accepted diagnostic cutoff for it. Avoid dismissing a symptomatic patient because several resting tests are normal.

Write a ranked explanation. Identify the strongest evidence, one important alternative, and the next test or management decision that would separate them. “Multifactorial” is useful only when the report names the contributing findings.

Noninvasive CPET can document limitation and suggest a circulatory component, but cannot reliably separate HFpEF from all causes of low fitness or peripheral limitation. Interpret it with the history, natriuretic peptides, echocardiography, and clinical probability. Obesity can suppress natriuretic peptides; AF can elevate them. Neither a low biomarker nor a normal resting filling-pressure estimate settles an exertional problem.

When a consequential question remains after noninvasive assessment, exercise echocardiography or invasive CPET may be appropriate in an experienced center. Invasive testing adds pressure, flow, and blood oxygen-content measurements during exercise, allowing differentiation of filling-pressure elevation, pulmonary vascular limitation, and selected flow/extraction abnormalities.

Audit transducer leveling, wedge position, respiratory convention, flow method, and blood-sampling timing before interpreting a borderline result. Large respiratory swings in obesity or obstructive lung disease can change classification. Do not interchange end-expiratory and respiratory-averaged thresholds without explaining the method.

A low arteriovenous oxygen difference can reflect impaired extraction, high flow, or measurement problems. “Preload failure” and peripheral extraction disorders require specialist evaluation rather than a label generated from one number. In suspected chronic thromboembolic disease, the diagnostic pathway can include ventilation–perfusion imaging; CPET cannot exclude or confirm chronic thromboembolic pulmonary hypertension.

In advanced heart failure, the 2024 ISHLT guidance includes peak V̇O₂ generally ≤12 mL/kg/min on beta blockers, or approximately ≤13–14 off beta blockers, as supportive considerations for transplant listing. Percent predicted and a V̇E/V̇CO₂ slope above 35 can add information, particularly in selected populations or submaximal tests. These are not standalone listing rules and must not be repurposed as “longevity cutoffs” for well adults.

For training, connect the prescription to the indication and the patient’s clinical stability. Threshold-related work, symptoms, and perceived exertion may be more useful than a watch’s generic heart-rate zones. Cardiac or pulmonary rehabilitation can provide monitored progression when indicated. A test does not replace an individualized exercise prescription.

For serial comparisons, keep modality, ramp, reference equation, averaging, medication timing, and oxygen conditions as consistent as possible. Compare symptoms and reason for stopping as well as the numerical peak. Record changes in body weight and hemoglobin. Distinguish better test familiarity from a treatment effect; a changed denominator or protocol can create an apparent improvement.

Improvement in function is meaningful even when it does not establish longer survival. CPET estimates neither biological age nor a personal expiration date.

1. Question and context: symptom, relevant diagnoses, medications, oxygen, and relevant baseline findings.
2. Method and validity: modality, protocol, duration, quality issues, averaging, reference equation, effort, and stop reason.
3. Capacity: peak uptake in all appropriate units, predicted comparison, workload, threshold and confidence.
4. Systems: circulation and rhythm; ventilation and mechanics; efficiency and oxygenation; symptoms and recovery.
5. Synthesis: dominant physiological pattern, important alternatives, and what the test cannot establish.
6. Action: a specific next decision, responsible clinician, and a follow-up plan appropriate to the clinical setting.

Prefer “Reduced aerobic capacity with marked ventilatory inefficiency and preserved estimated breathing reserve; investigate pulmonary vascular and left-heart causes” to “abnormal CPET.” Prefer “No convincing cardiopulmonary limit demonstrated before orthopedic termination” to “normal.” Include urgent findings in direct clinical communication rather than relying on a routine report.

Before signing

Could a different ramp, weight denominator, medication, or artifact explain the pattern? Did the test reproduce the symptom? Have you distinguished a measured finding from an inferred mechanism? What will the referring clinician do differently after reading the conclusion?