• Kay Choong See and David C Christiani.
• Division of Respiratory and Critical Care Medicine, Department of Medicine, National University Hospital, Singapore. kay_choong_see@nuhs.edu.sg
• Chest. 2013 Jan 1;143(1):107-16.

BackgroundElevated fractional excretion of exhaled nitric oxide (FENO) reflects airway inflammation, but few studies have established its normal values. This study aims to establish the normal values and thresholds for the clinical interpretation of FENO in the US general population.MethodsThirteen thousand two hundred seventy-five subjects aged 6 to 80 years sampled for the National Health and Nutrition Examination Survey (NHANES) 2007-2010 underwent interviews, physical examination, and FENO analysis at 50 mL/s using an online chemiluminescence device according to American Thoracic Society/European Respiratory Society guidelines. After excluding subjects with self-reported asthma and subjects with wheeze in the prior 12 months, prediction equations for the natural logarithm (ln) of FENO were constructed using age, sex, ethnicity, height, BMI, active/passive smoke exposure, and hay fever episodes as covariates.ResultsThe fifth to 95th percentile values of FENO were 3.5 to 36.5 parts per billion (ppb) for children < 12 years of age and 3.5 to 39 ppb for subjects 12 to 80 years of age. Using multiple linear regression, prediction equations explained only 10.3% to 15.7% of the variation in the general population. In the general population, 39% to 45% had ln(FENO) levels > 2 SD of the predicted means. When applied to the general population inclusive of subjects who reported asthma but who did not have attacks within the past year, nearly identical results were obtained.ConclusionsAssuming 95% of the healthy US general population had no clinically significant airway inflammation as assessed by FENO, values exceeding the 95th percentiles indicated abnormality and a high risk of airway inflammation. A large variation of normal FENO values existed in the general population, which was poorly predicted by multiple linear regression models.

21 keywords...

hide…