10.3 Evaluating DLCO Validity

Validity is a scored skill (II.C.9)

Domain II.C Evaluate task II.C.9 asks whether DLCO data are fit to report. A maneuver can look dramatic on the screen and still fail inspired-volume, timing, sample, or repeatability rules. High-cut RPFT items present pairs of trials, VI/VC ratios, breath-hold times, or clinical notes (smoking, strain) and expect you to accept, reject, repeat, or qualify the result.

Validity review assumes you already selected the protocol (10.1) and coached the sequence (10.2). Here you apply ATS/ERS single-breath DLCO standards concepts used throughout contemporary labs. Exact software flags vary by manufacturer; exam success comes from knowing what good looks like and which failures matter.

Acceptability criteria—major pillars

1. Inspired volume relative to vital capacity (VI / VC)

The patient must inspire a volume of test gas that is a large fraction of vital capacity. ATS/ERS-aligned teaching emphasizes:

• VI ≥ 90% of the largest measured VC (same visit) as a core acceptability target for adults.
• VI much below VC under-fills the lung, misrepresents alveolar volume available for exchange, and often yields non-comparable DLCO.

Worked example: Largest VC today = 4.00 L. Trial A VI = 3.70 L → 3.70/4.00 = 92.5% → inspired-volume OK. Trial B VI = 3.20 L → 80% → reject for incomplete inspiration even if breath-hold was exactly 10.0 s.

Exam trap: “Breath-hold was perfect, so low VI is fine.” Timing cannot rescue inadequate inspiration.

2. Breath-hold time (BHT)

• Target ~10 seconds.
• Acceptability is commonly framed as 10 ± 2 seconds (about 8–12 s).
• Too short → insufficient exchange time / unstable calculation assumptions.
• Too long → patient strain, leak risk, and departure from standardized conditions.

Know how your system defines BHT (Jones–Meade and related timing methods exist in standards literature). For the exam, emphasize ~10 s target and reject clearly out-of-window holds.

3. Sample collection quality (washout + alveolar sample)

Valid calculation needs alveolar gas that is not dead-space contaminated and not mixed with a fresh inspiration:

• Washout volume must be adequate to clear anatomic and apparatus dead space before sampling (adult default often ~0.75–1.0 L, reduced when VC is small per rules).
• Alveolar sample volume must be sufficient for analyzers and drawn from the appropriate exhaled segment.
• Sample during continuous expiration without inspiratory interruption.
• Watch for analyzer lag, water in the sample cell, or exhausted absorbers producing impossible tracer/CO relationships.

4. No major maneuver artifact

Even with VI and BHT in range, reject trials with:

• Mouth or nose leak
• Clear Valsalva or Müller physiology/coaching signs
• Cough that interrupts hold or sample
• Failure of valve switching (room air instead of test gas, or premature sample path)

Repeatability between acceptable maneuvers

Acceptability is per trial; repeatability is across trials. ATS/ERS DLCO standards concepts require more than one good maneuver:

• Obtain at least two acceptable tests.
• DLCO values must agree within 2 mL/min/mmHg (0.67 mmol/min/kPa) of each other. The 2017 ERS/ATS standard states this as a single absolute tolerance — it dropped the older 2005 “3 mL/min/mmHg or 10%” alternative. Do not fall back on a percentage escape hatch: two trials of 40.0 and 37.0 differ by 3.0, which is under 10% but fails the current criterion.
• Report the mean of acceptable, repeatable maneuvers (not the single highest “hero” trial that fails pairing).

Numeric repeatability examples

If two trials are each “pretty” but disagree widely, do not average junk—find a third acceptable trial or document limited reproducibility after max safe attempts.

VA and DLCO/VA checks (supporting validity sense)

Large swings in VA between trials often track inconsistent inspired volume, leak, or sample problems. Stable VI with unstable VA suggests technical artifact. Use these as clues while primary formal rules remain VI, BHT, sample, and DLCO repeatability.

Common failures—recognition table

High COHb and interpretation—validity vs reporting caveat

Elevated carboxyhemoglobin (recent smoking, exposure, many prior DLCO trials) reduces measured DLCO. Important distinctions for II.C.9:

• The maneuver may still meet VI, BHT, and sample rules—so it can be technically acceptable as a performance.
• The result may still be misleading for disease comparison if COHb is high and uncorrected/unnoted.
• RPFT behavior: document smoking/COHb, apply lab COHb adjustment when measured values and policy allow, and avoid declaring “new ILD progression” on a same-day post-cigarette study without technical commentary.
• COHb is therefore both a pre-test selection issue (10.1) and a post-test validity/reporting qualifier (II.C.9)—stems may place it in either frame.

Other interpretation-linked factors (usually beyond pure “accept/reject curve” but often bundled in evaluate items):

• Hb correction availability (anemia/polycythemia)
• FiO₂ / supplemental O₂ state during test
• Altitude and reference set consistency for serial tests

Session grading workflow (exam-ready)

1. After each trial, check VI/VC, BHT, leak/strain flags, and sample markers.
2. Label trial acceptable or reject with reason in the worksheet.
3. After ≥2 acceptable trials, test repeatability (within 2 mL/min/mmHg).
4. If not repeatable, add trials only within recovery interval and max trial limits.
5. If limits reached without a repeatable pair, report best available with quality grade / comments, or “unable to obtain reproducible DLCO,” per medical-director policy—never silently publish a single outlier as definitive.
6. Attach Hb, smoking/COHb, and O₂ notes for the interpreter.

Quality outcome examples

Example A — Reportable

• VC 3.50 L
• Trial 1: VI 3.25 L (93%), BHT 10.2 s, DLCO 24.0, no leak
• Trial 2: VI 3.30 L (94%), BHT 9.8 s, DLCO 23.4
• Difference 0.6 → mean DLCO 23.7 reportable; Hb applied per policy

Example B — Incomplete inspiration

• VC 3.50 L; VI 2.40 L (69%), BHT 10.0 s, DLCO 19.0
• Reject despite perfect clock time; reteach

Example C — Non-repeatable

• Two acceptable trials: DLCO 28 and 23 (diff 5)
• Need additional acceptable trial or limited study comment—do not average 28 and 23 as if repeatable

Example D — Smoking caveat

• Two repeatable trials, VI/BHT perfect, patient smoked 20 minutes prior, COHb high if measured
• Technically repeatable set with interpretation caveat; do not treat as clean baseline for drug-toxicity trending without note/correction pathway

Relationship to other tests in the visit

• Compare VA from DLCO with TLC from dilution/box when available; large unexplained discordance prompts technical review of both methods (volume chapter links).
• Severe obstruction may make SB inspiration and sample harder—validity may be limited even with excellent coaching; document.
• Do not “fix” a failed DLCO by rewriting BHT in the database without a real maneuver.

RPFT validity vignettes

Vignette 1: Software green-checks two trials with VI 78% and 80% of VC because BHT was 10 s. Action: override false confidence—reject on inspired volume; standards care about VI/VC, not only the timer.

Vignette 2: DLCO 25 then 18; both VI >90%, BHT OK; patient strained on second hold. Action: reject second as Valsalva-suspect; repeat after coaching; do not average 25 and 18.

Vignette 3: Only one acceptable trial after five attempts. Action: report single trial only with explicit quality limitation if policy allows, or unable—do not invent a second value.

Link to practice and later chapters

/practice/rpftPractice questions with detailed explanations