To examine the hypothesis that abnormalities on chest roentogram (CXR) will predict physiologic impairment on pulmonary function tests (PFTs) in patients with chronic obstructive pulmonary disease (COPD).
The CXRs of 108 consecutive patients with COPD were scored 0-4 with 1 point being awarded for the presence of each of the following: 1) flattening of the diaphragm, 2) increased retrosternal air, 3) hyperlucency, and 4) bullous changes. Demographic characteristics, absolute and percent of predicted (%) values of forced expiratory volume in one second (FEV1), forced vital capacity (FVC), total lung capacity (TLC), residual volume (RV), single breath carbon monoxide diffusing capacity (DLCO) and corrected for alveolar volume (DLCO/VA) were correlated with the CXR score. Data are presented as mean ± SD.
Radiographic signs of emphysema (CXR score ≥1) were seen in 67 patients (62%). Diaphragmatic flattening was the most common (n=46; 68%), followed by bullous disease (n=40; 59%) and retrosternal air (n=31; 46%). Patients with a CXR score ≥1 were older (63.5±10.3 vs. 58.1±11.1 years, p=0.01) and had more pack-years (48.5±20.8 vs. 36.7±18.9; p<0.01). Hyperinflation (increased TLC%), air-trapping (decreased FVC% and increased RV%) and gas exchange (decreased DLCO%) were all worse with increasing CXR scores as indicated in Table 1. A normal CXR score = 0 virtually excluded the possibility of hyperinflation (negative predictive value = 0.98). In contrast, a CXR score ≥1 was a good marker of functional hyperinflation (increased TLC% sensitivity = 0.96) and air trapping (increased RV/TLC% sensitivity = 0.81) while a score ≥2 was highly specific indicator of air trapping and decreased DLCO% (positive predictive values of 0.93 and 0.83 respectively).
The CXR score is an accurate predictor of the degree of functional impairment in patients with COPD. A normal CXR score obviates the need to measure lung volumes.
Disagreement between the CXR score and results of PFTs in the patient with COPD should prompt a search for technical errors in measurement.
Tarek Dernaika, None.Table 1CXR Score0 (n=41)1 (n=27)2 (n=22)≥ 3 (n=18)FEV1 %65±12a51±8b43±1445±15FVC %75±13.5c70±1364±1660±18TLC %96±10a111±12112±15118±21RV %125±25a179±46187±34200±58RV/TLC %44±8a54±14d60±965±8DLCO %80±26a73±29e58±2249±19a
= p < 0.05; 0 vs. 1, 2, 3;b
= p < 0.05; 1 vs. 2;c
= p < 0.05; 0 vs. 2, 3;d
= p < 0.05; 1 vs. 3;e
= p< 0.05; 1 vs. 2,3