Comparison of direct and indirect measurements of respiratory airflow: implications for hypopneas

The purpose of this study was to compare indirect methods for measuring respiratory airflow, such as temperature difference between inspired and expired air, thoracoabdominal movements, and nasal respiratory-airflow pressures-with a more direct measurement of minute ventilation using a head-out body plethysmograph. Measurements were obtained in healthy, awake, seated subjects during sequences of different levels of voluntary hypoventilations at 20 breaths/minute and analyzed to determine how well different methods could identify hypopneas (defined as reduction in minute ventilation by 50% or more). The results varied widely between different methods. Sensitivities ranged from 0 to 1, specificity ranged from 0.33 to 1, positive predictive values (PPV) ranged from 0 to 0.73, negative predictive values (NPV) ranged from 0.68 to 0.93. Cohen's kappa varied between 0 and 0.65 The poorest agreement was for the thermistor method, and the best agreement was obtained when a combination of thoraco-abdominal movements and nasal respiratory-airflow pressure was employed (sensitivity = 0.86, specificity = 0.83, PPV = 0.71, NPV = 0.92, Cohen's kappa = 0.65). We conclude that none of the indirect methods investigated, individually or in combination, proved adequate for identification of voluntary hypopneas in awake individuals.