Using the Lower Limit of Normal Instead of the Conventional Cutoff Values to Define Predictors of Pulmonary Function Impairment in Subjects With Chronic Heart Failure

Discussion

The current study showed that the definition used for pulmonary function impairment impacts the role of sex and smoking in pulmonary function in subjects with chronic heart failure with left ventricular systolic dysfunction. The lower limit of normal criterion identified an extra independent predictor of diffusion impairment compared with the 80% predicted value; in addition to BMI, pack-years, and V_A, female sex also turned out to be an independent predictor. A smoking history of ≥10 pack-years was a significant predictor of diffusion impairment and airway obstruction using the lower limit of normal criterion but not using the conventional cutoff values. However, making the conventional cutoff values more stringent by lowering the cutoff point yielded similar results as the lower limit of normal. Lower lung volumes were found in subjects with pulmonary congestion, cardiomegaly, and a history of coronary artery bypass grafting.

In the current study, the conventional cutoff values classified more subjects as having diffusion impairment and airway obstruction compared with the lower limit of normal. This is explained by the physiological decrease of the FEV₁/FVC ratio with age. The FEV₁ declines more rapidly with age than the FVC in normal subjects.²⁴ As a result, the fixed ratio of 0.7 that is traditionally used because of its simplicity may lead to overdiagnosis in the elderly and underdiagnosis in younger patients.¹⁸ Furthermore, the frequently used 80% predicted value has neither statistical nor physiological validity.^15–17 Limits of normal as the predicted ± 20% can only be accurate when the variance above and below the predicted regression line is proportional with the predicted value (ie, heteroscedastic: large variance with large values and small variance with small values). However, since this is not the case, because the scatter around the predicted regression line is constant (homoscedastic) in pulmonary function measurements, the 80% predicted rule of thumb may lead to false-positive diagnosis in the elderly and shorter individuals with smaller predicted values and underdiagnosis in younger and taller patients with larger predicted values.^15–17

Misdiagnosis of pulmonary function abnormalities by the conventional cutoff values may have interfered with the interpretation of prior research aiming to investigate the impact of heart failure and several clinical variables on pulmonary function. This may explain part of the inconsistencies across the studies. In fact, a smoking history of ≥10 pack-years was a significant predictor of diffusion impairment and airway obstruction using the lower limit of normal criterion but not using the conventional cutoff values. This implies that inclusion of subjects who are incorrectly labeled as having pulmonary dysfunction distorted the effect of smoking on pulmonary function. Indeed, lowering the conventional cutoff values to match the more stringent lower limit of normal and thus avoid overdiagnosis of diffusion impairment and airway obstruction in the elderly produced similar results as the lower limit of normal. Similarly, female sex was an independent predictor of diffusion impairment according to the lower limit of normal but not according to the 80% predicted value. However, decreasing the cutoff point to define diffusion impairment showed findings comparable to the lower limit of normal. On the other hand, by increasing the lower limit of normal to the 10th percentile, the association between female sex and diffusion impairment was lost (data not shown). This is explained by the fact that the lower ranges of diffusing capacity represented relatively more women than men. In summary, the lower limit of normal criterion identified more predictors of diffusion impairment and airway obstruction compared with conventional cutoff values in subjects with chronic heart failure with left ventricular systolic dysfunction. However, when conventional cutoff points were lowered to match the more stringent lower limit of normal criterion, the same effects were seen.

The effect of different definitions has also been put forward in the study of de Marco et al,²⁵ who have shown that the role of age, sex, former smoking, and low BMI on the development of COPD differs according to the definition used to define COPD. They suggested the need for a definition of COPD that is not exclusively based on spirometry.

Little is known about the clinical impact of different criteria of pulmonary dysfunction. Mannino and Diaz-Guzman²⁶ followed up the mortality data of a large number of subjects from the National Health and Nutrition Examination Survey III classified as normal, obstructed, or restricted using conventional cutoff values and the lower limit of normal. They found that subjects classified as normal using the lower limit of normal but obstructed or restricted using conventional cutoff values had a higher risk of mortality than normal subjects in up to 18 y of follow-up. This finding suggests that conventional criteria may identify at-risk patients who would have been missed using the lower limit of normal. This study was limited by the lack of post-bronchodilator pulmonary function test results, outcome parameters other than mortality, and the lack of comparison between subjects with mild airway obstruction according to conventional cutoff values (FEV₁/FVC <0.7 and FEV₁ ≥80% predicted) and normal subjects according to the lower limit of normal (FEV₁/FVC, FEV₁, and FVC ≥lower limit of normal). More longitudinal studies are warranted to determine which criterion is clinically more relevant in terms of morbidity (symptoms, exercise tolerance, health-related quality of life, hospitalization, use of health recourses) and mortality. Since we did not follow our subjects prospectively, it remains unknown whether pulmonary function impairment had prognostic implications in our study population and whether this is influenced by different definitions of pulmonary dysfunction. Several factors have been implied to play a role in the etiology of pulmonary function impairment in patients with heart failure, including the effects of heart failure itself on pulmonary function in addition to (previously undiagnosed) underlying pulmonary disease and confounding influences, such as smoking, coronary artery bypass grafting, and obesity.^4–7,13

Diffusion impairment has been thought to be related to the thickening of the alveolar-capillary membrane due to hydrostatic mechanical injury, interstitial edema, remodeling, and fibrosis.^1,2,4–6,27 Because heart transplantation does not affect or may even worsen pulmonary diffusing capacity despite an improvement in hemodynamic status and lung volumes,²⁸ it has been suggested that reduced diffusing capacity in chronic heart failure may be related to permanent damage to the alveolar-capillary membrane.² Other possible causes of diffusion impairment in heart failure include reduced lung and pulmonary capillary blood volumes, ventilation-perfusion mismatch, recurrent pulmonary emboli, smoking, and cardiopulmonary bypass.^4–6,13 The results of our study showed a higher NYHA class, smoking of ≥10 pack-years, pulmonary congestion, pleural effusion, and cardiomegaly to be associated with more impaired diffusing capacity, the latter 3 probably due to their negative effects on lung volume. Indeed, diffusing capacity corrected for V_A was not significantly different between the groups. Also, although V_A turned out to be an independent predictor of diffusion impairment, pulmonary congestion and cardiomegaly were not. In contrast to previous reports,^29–31 the use of angiotensin-converting enzyme inhibitors and aldosterone antagonists was not associated with increased diffusing capacity. Also, diffusing capacity was not significantly different between groups of subjects with or without a history of coronary artery bypass grafting. Independent predictors of diffusion impairment were BMI, pack-years (continuous variable), and V_A, whereas the role of sex and of having smoked ≥10 pack-years depended on the definition used to define diffusion impairment. Although the underlying mechanisms are not clear, women seemed to be more sensitive to the detrimental effects of heart failure on diffusing capacity. Sex differences in pulmonary function have been recognized before but not specifically in the heart failure population. Adult women have been reported to have lower resting lung diffusing capacity corrected for hemoglobin, smaller lung volumes, and lower maximal expiratory flows, even when corrected for age and standing height relative to men.³² It has been suggested that these sex differences in part can be explained by pulmonary structural differences (fewer total number of alveoli and smaller airway diameter relative to lung size) and hormonal influences in women.³² More research is needed regarding the influence of sex on pulmonary function in general and specifically in the heart failure population. The protective association between a higher BMI and less likelihood of having diffusion impairment has not, to our knowledge, been described before in subjects with chronic heart failure. However, some studies in healthy obese non-smokers have suggested that diffusing capacity may be increased in extremely obese subjects, probably as a result of the increase in blood volume.³³

Restriction has been linked to cardiomegaly, pleural effusion, respiratory muscle weakness, coronary artery bypass grafting, fibrosis from chronic congestion, and reduced lung compliance due to chronic vascular engorgement, interstitial/alveolar fluid accumulation, and chronic remodeling of the pulmonary vasculature due to elevated left atrial pressure.^{1,4–6,13,34,35} In line with expectations, we found lung volumes to be lower in subjects with pulmonary congestion, pleural effusion, cardiomegaly, and a history of coronary artery bypass grafting.

Airway obstruction has been attributed to alveolar fluid accumulation, bronchial mucosal swelling, peribronchial edema and fibrosis, squamous metaplasia of bronchial epithelial cells induced by transforming growth factor-β from the failing heart, geometric decrease in airway size from reduction in lung volume, abnormalities of autonomic control, neurohumoral bronchoconstriction, bronchial hyper-responsiveness, and smoking,^4–7,13,36 although results are not consistent.¹⁴ Our study showed a greater impaired FEV₁/VC ratio in subjects who had smoked ≥10 pack-years. Also, having smoked ≥10 pack-years was a significant predictor of airway obstruction, but this depended on the definition used to define airway obstruction. Although we excluded patients with known COPD or other obstructive lung disease, we cannot rule out the possibility that some of the subjects with airway obstruction had previously undiagnosed COPD, because most subjects were current or former smokers. In fact, 16% (lower limit of normal) to 24% (conventional cutoff values) of the subjects had post-bronchodilator airway obstruction after inhaling 400 μg salbutamol and 80 μg atrovent. These subjects had more symptoms of cough and sputum than those without post-bronchodilator airway obstruction (data not shown). Also, subjects who were current or former smokers tended more often to have post-bronchodilator airway obstruction than those who had never smoked (lower limit of normal: 19% vs 6%, P = .06; conventional cutoff values: 27% vs 11%, P = .05). However, hyperinflation, which has been found to be a valid indicator of true COPD in patients with congestive heart failure,³⁷ was not significantly different between groups of subjects with persistent airway obstruction after bronchodilation and those with reversible airway obstruction (data not shown). Although airway obstruction in heart failure has also been attributed to pulmonary congestion, the FEV₁/VC ratio was not significantly different between groups of subjects with and without pulmonary congestion in our study, which does not exclude the contribution of pulmonary congestion to small airway obstruction.

Importantly, airway obstruction may lead to hyperinflation of the lungs due to expiratory flow limitation and air trapping, which was found in almost half of our subjects with airway obstruction as defined by an increased residual volume/TLC ratio. This may contribute to symptoms of dyspnea, poor exercise tolerance, increased work of breathing and oxygen consumption, respiratory muscle dysfunction, and adverse impact on cardiac function by decreasing the preload.³⁸ Thus, irrespective of the causes, pulmonary function abnormalities associated with chronic heart failure may explain part of the symptoms and functional disability encountered in these subjects.^8–12 Moreover, pulmonary function impairment increases with the severity of heart failure,^9,11 provides important prognostic information,^39–42 and may ameliorate or normalize with several treatment modalities, such as pharmacologic and non-pharmacologic treatment of heart failure^2,28,43–46 and anti-obstructive therapy with bronchodilators.^47–53 Pulmonary function might thus be used as a guide for the evaluation of patients with chronic heart failure, with respect to severity of disease, prognosis, and response to treatment.

Some limitations of this study deserve further discussion. It is important to realize that these results may not be applicable to all patients with chronic heart failure, since we did not include subjects with preserved systolic function, who seem to have less impaired pulmonary function.⁵⁴ Furthermore, patients with more severe heart failure could have been underrepresented in this study because of an inability to participate, and thus pulmonary function abnormalities might have been underestimated. Finally, considering the relatively small number of subjects included in the current study, in particular women and non-smokers, further research is needed to confirm our results.