Increased understanding of the mechanisms and effects of acute respiratory failure has not been accompanied by more precise criteria by which the clinician can determine when intubation should be carried out and invasive positive-pressure ventilation (IPPV) instituted in a given patient. The indications traditionally offered in reviews and textbooks have tended to be either so broad as not to be very helpful in an individual case, or of questionable clinical relevance and too cumbersome for practical use. This review updates the indications for IPPV in adult patients with acute respiratory failure by examining available evidence from clinical trials and by considering new management alternatives that have become available in the last 20 years. Indications for IPPV based on specific threshold values for P(CO2) and pH or on various indices of arterial oxygenation have generally not been validated by clinical evidence, and it is unlikely that any cutoff value would be applicable to all patients or all categories of acute respiratory failure. Stated another way, there is probably no single value for arterial P(CO2), pH, or P(O2) that by itself constitutes an indication for IPPV. Compelling face validity justifies the use of IPPV in cases of apnea or when it appears certain that respiratory arrest is about to occur. However, dyspnea, tachypnea, or the subjective impression of respiratory distress are probably not in themselves justification for emergency intubation. It should be possible to avoid IPPV and its attendant complications in many cases of acute hypercapnic respiratory failure. In acute exacerbations of chronic obstructive pulmonary disease, noninvasive positive-pressure ventilation (NPPV) should be the initial ventilation approach unless the patient has one of several specific exclusion criteria such as cardiovascular instability or severely impaired mental status. It may also be possible to avoid intubation through the use of NPPV in certain immunocompromised patients with early acute hypoxemic respiratory failure. However, in other settings of acute hypoxemic respiratory failure, such as acute lung injury and acute respiratory distress syndrome, this has not been shown. The use of IPPV may improve outcomes in patients with severe cardiogenic shock. However, IPPV has not proven to be beneficial in traumatic brain injury and flail chest, in the absence of other indications.