In pneumonias, the short diagnostic results of non-invasive procedures lead frequently to use invasive techniques, among which we find pulmonary puncture aspiration (PPA). In this study, the profitability of PPA in the diagnosis of pneumonias is evaluated. One hundred and thirteen PPA were performed on 107 patients diagnosed of pneumonia. The PPA was performed with a 22-25 G needle without radioscopic control. Hemocultures were gathered in 104 cases, serological tests for productive agents of atypical pneumonia were applied in 50 cases, sputum test in 95 cases, and bronchofibroscopy with occluded telescopic brush in 25. The PPA had a specificity of 98 percent and a sensibility of 54% which rose to 73% in patients without antibiotic treatment prior to the puncture. Only 10 patients had complications with pneumothorax (9 percent) and four (3 percent) presented hemoptysic sputum. In the 53 patients with a positive PPA, knowledge of an etiological agent allowed the antibiotic treatment to be switched in 32 cases (60 percent). In addition, a decrease in the number of days of hospitalization was confirmed in the group of patients who received an etiological diagnosis while alive (p < 0.03). In conclusion, the PPA is a technique with excellent specificity, acceptable sensitivity, and a short rate of complications which allowed the treatment to be changed in 60 percent of the cases and reduced the number of days of hospitalization for patients who received a diagnosis. Thus, it constitutes a very useful technique in the etiological diagnosis of severe pulmonary infections.