15.2 DLCO & Exercise Test Reliability

DLCO Reliability Is More Than a Green Checkmark

After Domain II acceptability for single-breath DLCO, III.B.5 asks whether the reported diffusing capacity is reliable enough for clinical use. CO uptake numbers are sensitive to inspired volume, breath-hold timing, gas sample purity, and blood CO/hemoglobin state. An RPFT who only glances at “DLCO % predicted” without those checks will mislabel technical failure as interstitial disease—or miss a real diffusion defect hidden under a short breath-hold.

Core DLCO Reliability Checks

Inspired volume relative to VC (VI/VC)

The patient must inspire a large fraction of vital capacity so alveolar volume and CO distribution represent the lung being tested.

Exam stems often pair a “low DLCO” with a tracing showing small inspired volume and weak effort—reliability answer is retest, not pulmonary fibrosis narrative.

Breath-hold time (BHT)

BHT must match the method’s timing definition (ATS/ERS-aligned start/end rules used by the instrument and SOP). Too short or too long, or an unsteady hold with Valsalva/Müller straining, undermines the single-breath model.

Unreliable BHT patterns:

• Patient exhales early or inhales during “hold”
• Glottic leak (CO/tracer continuum unstable)
• Extreme pressures (strain) altering pulmonary blood volume
• Software BHT outside acceptable window with no override justification

Alveolar sample quality

DLCO depends on analyzing alveolar gas, not dead-space–heavy early gas or room-air contamination.

COHb and related confounders

Carboxyhemoglobin from smoking or CO exposure occupies Hb binding sites and alters measured DLCO. Reliability judgment includes:

• Documented smoking/vaping timing relative to test
• Measured COHb (when available) and appropriate correction/reporting practice per lab standards
• Anemia/polycythemia (Hb) affecting CO uptake—often corrected in reporting but still part of reliability/context
• Recent supplemental O2 affecting comparison to room-air reference conditions

A low DLCO in a patient who smoked immediately before testing is not automatically ILD until COHb and maneuver quality are addressed. Conversely, normal-looking DLCO with high COHb may still be clinically misleading without correction notes.

Replicate agreement

Within-session DLCO trials should cluster. Large scatter after several attempts means low reliability even if one value looks “pretty.” Prefer standards-aligned averaging rules only among acceptable trials.

Exercise & Field-Walk Reliability (III.B.6–7)

Exercise and walking tests produce distance, symptoms, SpO2, heart rate, and sometimes BP/ECG. Reliability fails when the protocol or sensors fail—even if the patient “tried hard.”

Protocol adherence

Oximeter reliability (motion & perfusion)

Pulse oximeters during walking are vulnerable to motion artifact, poor perfusion, ambient light, and probe placement.

Suspect SpO2 unreliability when:

• Pulse rate from oximeter disagrees with ECG/palpated rate
• SpO2 jumps erratically without physiologic pattern
• Cold fingers, nail coatings, or probe repeatedly reseated mid-test
• Patient clenches probe hand on walker/treadmill rail

Reliable practice: optimize probe site before start, confirm pulse agreement, treat wild SpO2 swings as artifact until proven, and document signal quality. Do not report a dramatic “desaturation” from a single motion-glitch nadir without waveform/pulse corroboration.

Practice walks and learning effect

A first 6MWT often underestimates true distance because of unfamiliarity. Standards commonly support a practice consideration or standardized repeat strategy. Reliability of “best distance” depends on whether the lab’s protocol for practice/repeat was followed and documented. Comparing today’s first-ever walk to last year’s third walk without noting learning effect is a reliability trap.

Course measurement errors

Distance is only as good as the course.

Exam answer when distance “falls 80 m” but the lab moved from a 30 m to an unmarked short hallway: question reliability of the comparison before declaring clinical worsening.

Home Testing Reliability Limitations (III.B.8)

Home spirometry and related remote testing improve access and serial sampling frequency, but reliability is constrained:

• Unsupervised effort — no real-time flow-volume coaching by an RPFT
• Device quality & calibration state — consumer/home sensors may lack full lab QC
• Environment — posture, nose clip use, distractions, illness day effects
• Maneuver selection — software may auto-pick peaks the patient cannot validate
• Context gaps — medication timing, smoking, recent exercise often incomplete

Reliability posture for the exam: Home data can trend and screen, but discordant or decision-critical values usually need laboratory confirmation. Reports should state unsupervised limitations rather than treat home FEV1 as identical in authority to a standards-aligned lab session.

Integrated Exam Scenarios

DLCO scenario: VI is 55% of measured VC, BHT short with leak flags, patient smoked 20 minutes prior, DLCO 45% predicted. Reliability: Do not lock in severe diffusion defect; correct process (inspiration, hold, COHb/timing) and retest.

Walk scenario: 6MWT SpO2 falls to 82% for two seconds while oximeter HR is 40 bpm and ECG shows 118 bpm; distance protocol correct. Reliability: SpO2 nadir is motion/perfusion artifact until a stable probe agrees with pulse; do not treat 82% as proven desaturation.

Home scenario: Home FEV1 drops 400 mL overnight; lab spirometry next day is repeatable and near baseline with good peaks. Reliability: Weight the supervised session higher; investigate home effort/device factors.

Workflow Checklist (III.B.5–8)

1. DLCO: VI/VC → BHT → sample/leak → replicate agreement → COHb/Hb/O2 context.
2. Exercise/walk: protocol clock & coaching → course length → practice-walk policy → SpO2 pulse agreement → symptom scores documented.
3. Home: label limitations; confirm critical decisions in lab when stakes are high.
4. If any pillar fails: retest or qualify—never silently convert technical failure into a disease label.

Mastering these reliability gates is what Domain III.B scores: not recalling the DLCO formula alone, but knowing when the formula’s inputs were never trustworthy.