[Accuracy of spirometry in the diagnosis of pulmonary restriction].

A restrictive ventilatory defect is characterized by a decreased total lung capacity (TLC). The objective of this study was to determine the accuracy of spirometry to detect pulmonary restriction in patients with or without airflow obstruction in the spirometry. Five hundred and twenty patients were included. Normal values for lung function were determined by using the 95% confidence interval (CI) with Morris reference equation for spirometry and European Respiratory Society equation for lung volume. Spirometries were considered obstructive when FEV1/FVC ratio was <70% and FEV1 was below 95%CI. In patients without obstruction in the spirometry (n = 357) sensitivity and specificity were 42.2% and 94.3% respectively, negative predictive value (NPP) was 86.6% and positive PV (PPV) was 65.2%. In patients with an obstructive spirometry (n = 66) sensitivity increased to 75.8% but specificity decreased to 65.9%. PPV was only 57.8% and NPV 81.5%. Patients showing obstruction in the spirometry and false positives of a low FVC (n = 22) had similar values of FVC (57.36 +/- 13.45 vs. 58.82 +/- 8.71%, p = 0.6451), FEV1 (44.73 +/- 19.24 vs. 44.0 +/- 13.08%, p = 0.8745) and DLCO (67.50 +/-27.23 vs. 77.00 +/-16.00%, p = 0.1299) than true positives. Residual volume (RV) (125.72 +/- 64. vs. 77.96 +/- 29.98%, p = 0.0011) and RV/ TLC ratio (56.89 +/- 12.82 vs. 38.43 +/- 13.07%, p = <0.0001) were significantly higher. We conclude that a decreased FVC or VC cannot be considered evidence of ventilatory restriction in the presence of airflow obstruction on spirometry. Diagnosis of "mixed defects" by spirometry is inaccurate and should be avoided without the measurement of lung volumes.