The objective of this study was to evaluate the risk factors associated with mortality in interstitial lung disease patients. We performed a retrospective study of 722 consecutive patients submitted for lung biopsy during the 1986-1990 period. Twenty-two (3%) died within the 30 days following surgery. Forty-four patients who survived after the surgery for the same time span were randomly chosen as control group. Dyspnea at rest was present in 18/44 of surviving group (SG) and in 18/22 of the nonsurviving group (NSG) (OR 6.5, 95% CI 1.8-22.4,p = .001). Systemic diseases (i.e., diabetes, systemic arterial hypertension)were mainly present in the NSG (OR 7.2, 95% CI 2.3-22.8, p < .001). The SG displayed significantly less respiratory insufficiency with a PaO2 of 52.2 + 8.4 versus 38.5 i 9.4 mm Hg, and PaCO2 of 28.8 i 4.5 versus 38.5 +/- 9.2 mm Hg, respectively (p < .001). Likewise, the SG exhibited a PaCO2/PaO2 ratio of 0.5 - 0.1, while in the NSG it was of 1 +/- 0.4 (p < .001), showing a sensitivity of 84% and specificity of 93% for mortality. Multiple logistic regression analysis for these variables showed that log likelihood was still significant for PaCO2 > 34 mm Hg, PaO2 <48 mm Hg, and comorbid diseases. Logistic regression analysis of these three variables showed the greatest sensitivity and specificity (84 and 750/0,respectively) for prediction of mortality. However, the strongest association was found when PaCO2/PaO2 ratio was analyzed alone (OR 21,073,CI 95% 28-15,946,357, p < .005). These data suggest that PaCO2/PaO2 ratio appears to be a predictor of mortality in this subset of patients. Its prospective use has reduced early mortality after surgery less than 1% in the last decade.