How do I interpret a DLCO (diffusing capacity of the lung for carbon monoxide) result?

DLCO Interpretation

Interpret DLCO by first assessing test quality, then comparing the result to the 5th percentile lower limit of normal (not 80% predicted) using reference equations matched to the patient's measured height, age, sex, and ethnicity, and finally integrating the result with spirometry, lung volumes, and clinical context to identify the underlying physiological pattern. 1

Step 1: Assess Technical Quality First

Before relying on any numerical DLCO value, review the raw data and ensure the test meets acceptability and reproducibility criteria. 1 This is the most frequent source of interpretation error. 1 A technically inadequate test may still provide useful clinical information, but you must document the direction and magnitude of potential errors. 1

Step 2: Select Appropriate Reference Values

Use reference equations that match your patient's characteristics:

• Measure height with a calibrated stadiometer at the time of testing—never use self-reported height. 1, 2
• Select equations based on the patient's age, sex, measured height, and ethnicity. 1, 2
• Ensure all parameters (DLCO, alveolar volume [VA], transfer coefficient [KCO]) come from the same reference source to maintain internal consistency. 3, 2

Critical caveat: Inter-laboratory variability for DLCO is much larger than for spirometry, making reference value selection more problematic. 3 Laboratory directors should ideally validate their chosen equations against a sample of healthy subjects tested in their own lab. 3, 2

Step 3: Apply Ethnicity Adjustments

• For Black patients, apply a 0.88 correction factor to predicted DLCO values (similar to the 12% reduction used for lung volumes). 1, 2, 4
• For Asian-American patients, use a 0.94 correction factor. 1
• Without these adjustments, up to 50% of healthy Black individuals may be incorrectly labeled as abnormal using standard Caucasian-derived equations. 4

Step 4: Define Normal vs. Abnormal

Use the 5th percentile of the reference population as the lower limit of normal (z-score ≤ -1.64), not a fixed 80% predicted cutoff. 1, 2 The 80% threshold leads to substantial misclassification, particularly in older adults. 1

• Values below the 5th percentile are abnormal. 2
• Values close to thresholds carry the highest risk of misclassification and may warrant repeat testing or additional studies. 1

Step 5: Adjust for Physiological Variables

Measure and adjust for:

• Hemoglobin concentration: Anemia artificially lowers DLCO (fewer binding sites for CO), while polycythemia raises it. 5, 2 This adjustment is particularly important in autoimmune diseases and when monitoring for drug toxicity. 5
• Carboxyhemoglobin: Elevated levels (e.g., in smokers) reduce available hemoglobin and falsely lower DLCO. 3, 5, 2
• Altitude: Higher altitude affects gas transfer. 3, 2

Step 6: Interpret in Context of Lung Volumes

Never interpret DLCO in isolation. 2 The relationship between DLCO and alveolar volume (VA) provides critical diagnostic information:

When VA is Reduced (Restrictive Pattern):

• DLCO decreases more than DLCO/VA in true interstitial lung diseases. 6
• Standard reference values may be inappropriate—consider correcting theoretical formulas for the patient's actual VA. 6
• Confirm restriction with total lung capacity (TLC) below the 5th percentile; VA from single-breath DLCO systematically underestimates TLC by up to 3 liters in severe obstruction. 1

When VA is Increased (Hyperinflation):

• DLCO/VA decreases more than absolute DLCO in emphysema. 6
• In patients with overinflated lungs, the percentage of predicted DLCO is more increased than DLCO/VA. 6

When Both are Equally Reduced:

• Suggests emphysema. 6

Step 7: Recognize Patterns by Clinical Condition

Low DLCO (<60% predicted):

• High-risk marker: Associated with 25% mortality and 40% pulmonary morbidity in patients undergoing lung resection. 1
• Common causes: Interstitial lung disease, emphysema, pulmonary vascular disease, anemia. 1

High DLCO (>140% predicted):

• Most commonly associated with obesity (62% of cases), asthma, or both. 7
• Also seen with large lung volumes, increased body surface area, and higher body mass index. 7
• Polycythemia, hemoptysis, and left-to-right shunts are uncommon causes. 7
• Important pitfall: A condition that typically reduces DLCO may deceptively normalize it in obese or asthmatic patients. 7

Normal DLCO at Rest with Unexplained Symptoms:

• Consider cardiopulmonary exercise testing (CPET): In sarcoidosis, 70% of patients with normal DLCO show moderate-to-severe gas exchange impairment during maximal exercise. 8
• Normal resting DLCO is an inappropriate predictor of exercise-induced gas exchange abnormalities. 8

Step 8: Age-Specific Considerations

In patients aged 65-85 years, reference equations derived from middle-aged adults significantly underestimate true DLCO:

• Underestimation ranges from 2-37% in men and 2-22% in women. 9
• Use age-appropriate reference equations to avoid falsely labeling healthy elderly individuals as abnormal. 9

Common Pitfalls to Avoid

1. Relying on computer interpretations without reviewing test quality and raw data. 1
2. Using fixed 80% predicted cutoffs instead of the 5th percentile lower limit of normal. 1, 2
3. Failing to adjust for hemoglobin and carboxyhemoglobin, especially when monitoring drug toxicity. 1, 5
4. Not measuring TLC when restrictive physiology is suspected based on reduced VC alone—only ~50% of low-VC cases have low TLC. 1
5. Interpreting multiple parameters simultaneously inflates false-positive rates: examining 14 parameters yields a 24% abnormality rate in healthy subjects. 1
6. Changing interpretation strategies within a laboratory can create the illusion of disease progression when none exists. 1
7. Ignoring ethnicity adjustments, leading to up to 50% false-positive rates in Black patients. 4

Severity Grading

When DLCO is reduced, grade severity primarily by FEV1 % predicted (not DLCO itself) in the context of obstructive or mixed defects: 1

• Mild: >70%
• Moderate: 60-69%
• Moderately severe: 50-59%
• Severe: 35-49%
• Very severe: <35%

However, DLCO <60% independently predicts higher mortality and morbidity, regardless of FEV1. 1