Hennessy and colleagues' experience of no survivors out of 21 mechanically ventilated patients is consistent with our findings and those reported in the literature. We note that the upper bound of the 95% CI in this small sample allows for a 16% survival rate in mechanically ventilated patients. It would be problematic to draw conclusions about the relation (or lack of a relation) between timing of respiratory failure after bone marrow transplantation and survival in a sample of 21 patients.

In our report we proposed guidelines for limiting life support in selected bone marrow transplant recipients on the basis of three principles: 1) prognostic indicators with high specificity and narrow CIs; 2) a process through which patients and their surrogates are informed of the risks of respiratory complications, the outcomes of mechanical ventilation, and the decision-making guidelines; and 3) an ongoing educational and quality improvement process to evaluate end-of-life decisions.

Most patients who eventually die meet one of the criteria for not surviving by day 4 of mechanical ventilation. The prognosis of patients who do not fulfill any of the criteria is 13.5%. For patients who consent, we therefore believe that a trial of mechanical ventilation is justified and that life support should be promptly withdrawn if the patient does not respond. Lack of response is manifested by hepatic and renal failure, lung injury, or hemodynamic instability requiring vasopressors.