Surya Bhatt, M.D.

Considerable controversy exists on the best methods to diagnose airflow obstruction. Most of the discrete ratio thresholds used to define airflow obstruction are insensitive to early and mild disease. Although several efforts have been made to develop other measures for detection of airflow obstruction, these have not resulted in clinically useful measures with sufficient discrimination from normal. Additionally, demographics like age can influence currently-used airflow obstruction measures, complicating diagnostic threshold selection for airflow obstruction.

To combat this, a research team led by Surya Bhatt, M.D. proposed a new airflow obstruction metric Parameter D. In the current work, Dr. Bhatt and team evaluated the influence of demographics on Parameter D and showed that this new measure is minimally influenced by most demographics. In addition, the team reported new diagnostic thresholds for parameter D based on normal population that could aid in the early identification of more individuals with airflow obstruction.

Bhatt’s research team analyzed the spirometry data of normal subjects enrolled in the 2007-8, 2009-10 and 2011-12 NHANES cohorts and calculated Parameter D using the expiratory volume-time curve. Relationships between demographics and lung function (FEV1, FEV1/FVC, and Parameter D) were assessed using generalized linear models in NHANES and UK Biobank. Based on concordance between the lower-limit-of-normal (LLN) for FEV1/FVC and the Parameter D threshold, four groups were found: Normal (no airflow obstruction by either criterion), D+COPD (positive by Parameter D only), D-COPD (positive by LLN only), and COPD (positive by both criteria). Any associations with structural lung disease, exacerbations, and mortality were assessed using multivariable analyses.

The variance of Parameter D explained by each demographic feature is very low, in contrast to FEV1. FEV1 is significantly influenced by body size and its correlates of height, sex, and race. Age-related loss of lung elasticity also has a negative impact on FEV1. In contrast, it is plausible that in the case of Parameter D, the rate of volume change reflects proportions of the preceding segments of the curve, meaning the impact of lung size is negligible. The FEV1/FVC ratio, by including lung size in the denominator, is minimally influenced by height but decreases substantially with advancing age. Parameter D, in contrast, is not influenced by age.

Based on the frequency distribution of Parameter D in a representative healthy community dwelling population, the research team also discovered a threshold that results in the identification of additional individuals with a substantial amount of structural lung disease and respiratory symptoms. Over 25% of those with airflow obstruction identified by Parameter D alone were found to have airflow obstruction by traditional criteria 5 years later. When this is compared with 8% of normal, this suggests that this metric can also find airflow obstruction earlier.

These advantages are especially noteworthy at a time when the inclusion of demographics such as race and ethnicity in reference equations is more present than it has ever been. With race/ethnicity based equations, there is potential for misdiagnosis in the minority populations with implications for delayed diagnosis, withheld treatments, and lower access to disability benefits. The work has been conducted in collaboration with Dr. Sandeep Bodduluri (UAB Pulmonary), Dr. Arie Nakhmani (UAB Electrical Engineering), and led by Dr. Surya Bhatt (UAB Pulmonary).