Severity Classification Using New Guidelines for Spirometry Interpretation: Links to Outcomes - COPD: Peer Perspectives (2024)

According to a new study, severity classification of airflow obstruction using race-neutral reference equations and z scores was associated with risk of exacerbation and all-cause hospitalization, supporting the recent changes in guidelines for spirometry interpretation.¹

Recommendations for the interpretation of spirometry severity classifications have recently been updated. The European Respiratory Society and the American Thoracic Society recommended changing classification systems from percent predicted forced expiratory volume in 1 second (FEV1) with 5 levels to z scores with 3 levels in 2022.² In addition, in 2023, the American Thoracic Society recommended ending race and ethnicity in the interpretation of pulmonary function tests.³

Some concerns have been raised about these changes.

“Although the recent shift in recommendations to using race-neutral equations has received a lot of attention, it hasn’t been well understood how that shift may variably impact patients by whether clinicians are using percent predicted values or z scores,” said J. Henry Brems, MD, of the Division of Pulmonary and Critical Care Medicine and the Berman Institute of Bioethics, Johns Hopkins University, Baltimore, MD, in an interview with MedPage Today.

“Perhaps more importantly, a key concern regarding the shift to race-neutral equations has been whether these equations could lead us to misclassify disease severity,” said Dr. Brems. “For example, critics of the change have argued that we may now overestimate disease severity among Black patients and underestimate it among White patients.”

Dr. Brems explained, “We sought to address that concern in this study by evaluating whether the patients who are now classified as higher or lower severity with race-neutral equations actually have a corresponding higher or lower risk of disease-related outcomes.”

Study design

The design was a retrospective cohort study. Data were obtained from Johns Hopkins Health System electronic health records of adults with chronic obstructive pulmonary disease (COPD) who had spirometry findings between January 2015 and January 2023. Inclusion criteria were a diagnosis of COPD or obstruction, defined as an FEV1 to forced vital capacity ratio < .70, and Black or White race.¹

Data from the first spirometry report in the records were used. The researchers calculated FEV1 percent predicted and z score for each patient. Race-specific equations were from the Global Lung Function Initiative 2012 equations, with race-neutral equations from the Global Lung Function Initiative 2022 equations.^4,5

Percent predicted values were classified as mild (≥ 70%), moderate (70% to 50%), and severe (< 50%). For z scores, classification thresholds were mild (≥ –2.5), moderate (–2.5 to –4), and severe (< –4).¹

The primary outcome was COPD exacerbation within 1 year of spirometry. An exacerbation included an emergency department visit, admission for observation, or inpatient admission with a primary or secondary International Classification of Diseases, Tenth Revision, code of J44.1 for COPD with acute exacerbation. The second primary outcome was all-cause hospitalization.¹

Patient characteristics

The study included 13,324 patients with COPD or obstruction, with 9232 White patients and 4092 Black patients. The mean [SD] age of White patients was 65.7 [12.7] years, and the mean age [SD] of Black patients was significantly younger at 61.1 [11.7] years (P < .001). A lower proportion of White patients were female (53.0%) compared with Black patients (61.5%; P < .001).¹

FEV1 was also significantly lower in Black patients compared with White patients. The average absolute FEV1 was 1.82 L for White patients and 1.61 L for Black patients (P < .001).¹

Effect of race-neutral reference equations on classification

Using race-neutral equations increased the average FEV1 percent predicted and FEV1 z scores for White patients and lowered them for Black patients compared with the 2012 race-specific equations. For White patients, FEV1 percent predicted increased by an average of 2.8 percentage points, and FEV1 z scores increased by an average of 0.29 points. For Black patients, FEV1 percent predicted decreased by an average of 7.1 percentage points, and FEV1 z scores decreased by an average of 0.41 points. In 10.7% of patients, the change in FEV1 percent predicted and FEV1 z scores were not in the same direction.¹

When FEV1 percent predicted was used to classify disease severity, race-neutral equations affected the classification of Black patients more often than White patients. A total of 6.1% of White patients had a change in classification, most to less severe, while 20.2% of Black patients had a change in classification, all to more severe (P < .001).¹

In contrast, when using FEV1 z scores to classify disease severity, the proportion of patients with different classifications after using race-neutral equations was not different between racial groups. A total of 12.3% of White patients and 12.6% of Black patients had a change in classification (P = .68).¹

Classification schemes: association with clinical risk

In the 12 months following the spirometry assessment, 3.0% of patients had an exacerbation, and 32.7% were hospitalized for any cause.¹

The researchers used logistic regression models to analyze the association between outcomes and classification when using race-neutral equations. This analysis showed that classification using z score thresholds better corresponded to clinical risk than FEV1 percent predicted thresholds.¹

For FEV1 percent predicted thresholds, a lower severity classification with race-neutral equations was associated with a decreased risk for exacerbation compared with no change in classification (odds ratio [OR], 0.49; 95% confidence interval [CI], 0.28-0.87). However, a higher severity classification was not associated with an increased risk for exacerbations compared with no change in classification (OR, 1.08; 95% CI, 0.61-1.93).¹

For FEV1 z score thresholds, a change in classification was appropriately associated with the risk of exacerbations. A lower severity classification was associated with a decreased risk for exacerbation (OR, 0.67; 95% CI, 0.50-0.90), and a higher severity classification was associated with an increased risk for exacerbation (OR, 2.34; 95% CI, 1.51-3.63) when using z score thresholds compared with no change in classification.¹

Using either FEV1 percent predicted thresholds or FEV1 z score thresholds for classification, a change to a higher disease severity classification was associated with an increased risk for all-cause hospitalizations with no change in classification. Lower disease severity was also associated with a lower risk for all-cause hospitalizations for both FEV1 percent predicted and FEV1 z score thresholds.¹

Limitations and support for the new guidelines

One study limitation is the lack of analysis of other races and ethnicities. “There also remains an important need to understand the impact of transitioning to race-neutral equations for patients of races other than White or Black,” said Dr. Brems.

Other limitations were generalizability to other populations and potential missing exacerbations and hospitalizations that occurred outside the Johns Hopkins Health System or were incorrectly coded.¹

The researchers concluded that when using race-neutral equations, z score thresholds changed the severity classification for the same proportion of Black and White patients and were associated with clinical risk of exacerbations and all-cause hospitalizations.¹

The researchers did note in their paper that percent predicted values may still be valuable for diagnostic thresholds, and “because they may be more intuitive than z scores, have ongoing usefulness for both patients and clinicians less familiar with spirometry interpretation.”¹

However, “physicians should feel more confident in following the recent guideline recommendations both to use race-neutral reference equations and to use z scores for assessing severity,” said Dr. Brems. “Together, these changes seem to accurately classify disease risk for patients with COPD.”

“Further work should aim to understand the impact of the transition to both race-neutral equations and z scores for other disease states and outcomes, as we only investigated exacerbations and hospitalizations among patients with COPD.”

Published: May 28, 2024