How do you interpret pulmonary function test (PFT) results based on the z-score for an adult patient with respiratory symptoms and a history of lung disease, such as asthma or Chronic Obstructive Pulmonary Disease (COPD)?

Interpreting Pulmonary Function Tests Using Z-Scores

Pulmonary function test results should be interpreted by expressing values as z-scores (standard deviations from the predicted mean), with abnormality defined as a z-score below -1.645 (the lower limit of normal at the 5th percentile), rather than using arbitrary percentage cutoffs like 80% of predicted. 1

Understanding Z-Scores in PFT Interpretation

What Z-Scores Represent

• Z-scores indicate how many standard deviations a patient's measured value falls above or below the predicted mean for a reference population matched by age, height, sex, and ethnicity 1
• The formula is: z-score = (observed value - predicted mean) / residual standard deviation of the reference population 1
• A z-score of 0 represents the population mean, while negative values indicate below-average function 1
• Z-scores follow a normal distribution, allowing probability-based interpretation of how likely a result is to occur in healthy individuals 1

Advantages Over Percent Predicted

• Z-scores are superior to percent predicted values because they account for the natural variability in the reference population and avoid arbitrary fixed cutoffs 1
• Percent predicted values can be misleading, particularly at extremes of age or height, where the coefficient of variation changes 1
• Z-scores allow direct comparison across different pulmonary function parameters (FEV1, FVC, DLCO) and facilitate tracking changes over time 1

Systematic Interpretation Algorithm Using Z-Scores

Step 1: Assess Test Quality First

• Never interpret numerical results without first reviewing technical adequacy of the maneuvers 2, 3
• Verify that acceptability and reproducibility criteria are met according to ATS/ERS standards 1, 2
• Poor technique produces unreliable results regardless of z-score values 3

Step 2: Define Normal vs. Abnormal

• Use the lower limit of normal (LLN) defined as the 5th percentile, corresponding to a z-score of -1.645 1, 2
• Values with z-scores ≥ -1.645 are considered normal 1
• Values with z-scores < -1.645 are considered abnormal 1
• This approach is more accurate than using fixed percentages like 80% predicted, which can misclassify patients 2, 3

Step 3: Identify the Physiologic Pattern

• First evaluate FEV1/FVC ratio (not FEV1 alone) to determine if obstruction is present 3, 4
• If FEV1/FVC z-score < -1.645 → Obstructive pattern 3
• If FEV1/FVC z-score ≥ -1.645 and FVC z-score < -1.645 → Possible restrictive pattern, but must measure TLC to confirm 3, 4
• Never diagnose restriction based on spirometry alone—reduced FVC has poor positive predictive value without TLC measurement 3, 4
• If TLC z-score < -1.645 → Confirmed restrictive pattern 3

Step 4: Grade Severity Using FEV1 Z-Score

Once the pattern is identified, severity classification for obstructive, restrictive, and mixed defects is based on FEV1 percent predicted (not z-score directly), but z-scores help track progression 2:

• Mild: FEV1 >70% predicted (z-score approximately -1.0 to -1.645)
• Moderate: FEV1 60-69% predicted (z-score approximately -1.6 to -2.3)
• Moderately severe: FEV1 50-59% predicted (z-score approximately -2.3 to -3.0)
• Severe: FEV1 35-49% predicted (z-score approximately -3.0 to -4.0)
• Very severe: FEV1 <35% predicted (z-score < -4.0) 2

Step 5: Assess DLCO Using Z-Scores

• DLCO z-score < -1.645 indicates abnormal gas transfer 2
• DLCO <60% predicted (z-score typically < -2.5) is associated with significantly higher mortality and pulmonary morbidity, particularly in preoperative lung resection candidates 2
• Always adjust DLCO for hemoglobin and carboxyhemoglobin levels 2

Clinical Application in Asthma and COPD

For COPD Patients

• FEV1 z-score correlates with symptom severity and prognosis according to the European Respiratory Society 2, 3
• Serial z-scores allow objective tracking of disease progression independent of aging effects 1
• Inspiratory capacity and DLCO z-scores are important mortality predictors beyond FEV1 alone 3

For Asthma Patients

• Bronchodilator response is significant if FEV1 or FVC increases by ≥12% and ≥200 mL from baseline 4
• Z-scores help identify variability over time—changes >12% in FEV1 support asthma diagnosis 3
• Normalization of z-scores post-bronchodilator (z-score returning to ≥ -1.645) indicates reversible obstruction 3

Critical Pitfalls to Avoid

Common Interpretation Errors

• Never diagnose obstruction based on reduced FEV1 z-score alone—always check the FEV1/FVC ratio first 3, 4
• Never confirm restrictive disease without measuring TLC, as reduced FVC on spirometry has poor positive predictive value 3, 4
• Never use the FEV1/FVC ratio to grade severity—use FEV1 percent predicted instead 2, 3
• Never rely solely on computer-generated interpretations without personally reviewing the flow-volume loop 2, 3

Reference Equation Selection

• Ensure the reference equations used match the patient's demographics (age, height, sex, ethnicity) 1
• All spirometric parameters must come from the same reference source to ensure consistency 3
• Be aware that race-specific equations may artificially lower predicted values in Black individuals, potentially masking clinically important disease 3

Tracking Changes Over Time

• Z-scores are particularly valuable for longitudinal monitoring because they account for normal age-related decline in lung function 1
• A change in z-score of >1.0 (approximately 10-15% decline in FEV1) over time suggests clinically significant deterioration 1
• Ensure interpretation consistency within your laboratory to avoid inferring patient changes that are actually due to different interpretation approaches 2

Special Considerations for Preschool Children

• Z-scores are the preferred method for expressing results in children aged 2-6 years 1
• Variability measurements should not be extrapolated from healthy children to those with disease 1
• Within-subject variability assessments need at least 30 subjects of similar age and diagnostic category to establish reproducibility 1