The occurrence of acute respiratory failure (ARF) is often seen by oncologists as a terminal phase of the disease, this view being based on studies reporting limited survival at considerable costs in such patients. A large proportion of cancer patients with severe respiratory failure are denied admission to an ICU because intensive care specialists are aware that intubation and mechanical ventilation are both strong predictors of mortality in critically ill cancer patients. This holds particularly true in the subset of patients who are not receiving chemotherapy or radiotherapy because of the advanced stage of their disease, and who are also not affected by an episode of ARF, related to a reversible cause.

These patients often receive oxygen therapy and morphine in an attempt to improve oxygenation and/or relieve the ensuing dyspnea.

Non-invasive mechanical ventilation (NIV) is now the first line treatment of ARF in selected populations (e.g., those with COPD) and has been used sporadically as a potential treatment of acute respiratory failure in patients with a "do-not-intubate" order. The International Consensus Conference on Intensive Care Medicine stated that "the use of NIV may be justified in selected patients who are "not to be intubated" and may provide patient comfort and facilitate physician-patient interaction." "Early" NIV has been successfully used so far in cancer patients only to prevent intubation among those with hematologic malignancies, while a pilot study has assessed the feasibility of NIV also as a "palliative" treatment of end-stage solid cancer patients. So far we are lacking data about the "pure palliative" effects of NIV,in patients with end-stage solid cancer.

The aim of this multicenter randomised study will be to evaluate on a large scale the feasibility, clinical efficacy and impact on quality of life and dyspnea of NIV versus standard medical in patients with respiratory failure, not related to a reversible cause, and solid cancer needing palliative care treatment.

Rationale:

The occurrence of ARF is often seen by oncologists as a terminal phase of the disease, this view being based on studies reporting limited survival at considerable costs in such patients (1, 2, 3, 4, 5, 6). A large proportion of cancer patients with severe respiratory failure are denied admission to an ICU because intensive care specialists are aware that intubation and mechanical ventilation are both strong predictors of mortality in critically ill cancer patients (1, 3, 5, 7). This holds particularly true in the subset of patients who are not receiving chemotherapy or radiotherapy because of the advanced stage of their disease, and who are also not affected by an episode of ARF, related to a reversible cause.

These patients often receive oxygen therapy and morphine in an attempt to improve oxygenation and/or relief the ensuing dyspnea.

Non-invasive mechanical ventilation (NIV) is now the first line treatment of ARF in selected populations (e.g. those with COPD) (8) and has been used sporadically as a potential treatment of acute respiratory failure in patients with a "do-not-intubate" order (9). The International Consensus Conference on Intensive Care Medicine stated that "the use of NIV may be justified in selected patients who are "not to be intubated" and may provide patient comfort and facilitate physician-patient interaction" (8). "Early" NIV has been successfully used so far in cancer patients only to prevent intubation among those with hematologic malignancies (10), while a pilot study has assessed the feasibility of NIV also as a "palliative" treatment of end-stage solid cancer patients (11). So far we are lacking data about the "pure palliative" effects of NIV,in patients with end-stage solid cancer.

Aim. The aim of this multicenter randomised study will be to evaluate on a large scale the feasibility, clinical efficacy and impact on quality of life and dyspnea of NIV vs standard medical in patients with respiratory failure, not related to a reversible cause, and solid cancer needing palliative care treatment.

Study design Multicenter randomised study. The sample size has been estimated on about 50 patients for each group (i.e. hypercapnic and hypoxic ARF, with different randomisation) , TOTAL= 200 pts based on the survival rate of cancer patients with acute respiratory failure and DNR order (18%) (2) versus that of patients treated with NIV in a pilot study (57%) (14).

You will receive separate random sequence according to the degree of PaCO2. Methods Consecutive solid cancer patients with ARF, in which palliative treatment is the only one indicated, and: 1) have signed or expressed a DNI order or 2) judged by the attending physician having < 6 months or life expectancy (Mc Crabe index= class 3-4) or PPI > 5 (Palliative Prognostic Index) and not deserve to be intubated by a "blind" ICU physician.

A brief trial lasting < 5 minutes will be performed before randomization using either NIV or Venturi mask to assess the willing of the patient to be enroll in the protocol (since they are not familiar with both techniques).

Major criteria for enrollment into the study were one of the following: 1) PaO2/FiO2 ratio < 250 + one of the two following: 1) dyspnea with recruitment of the accessory muscles and/or abdominal muscles recruitment and 2) respiratory rate > 30 b/ min. Hypercapnia per se is not a criteria of inclusion, but it is not an exclusion criteria if chronic.

Exclusion criteria were: potentially reversible causes of exacerbation such as (CPE, pneumonia or exacerbation of chronic pulmonary disorders) coma, refusal of treatment, inability to protect the airways, an agitated or uncooperative patient, anatomical abnormalities interfering with mask fit, uncontrolled cardiac ischemia or arrhythmias, failure of more than two organs.

NIV settings. The positive end-expiratory pressure initially set at 5 cm H2O and could be increased by 1 cm H2O until a brisk increase in oxygen saturation (SaO2) is observed, whereas the inspiratory pressure support is initially set at 10 cm H2O and then increased in increments of 2 cm H2O up to the maximum tolerated.

For patients with hypercapnic respiratory failure the inspiratory pressure is adjusted according to the patient's tolerance, with the external PEEP not exceeding 6 cm H2O.

In any case both settings are aimed to achieve a respiratory rate < 25 breaths/min, no evident sign of wasted respiratory efforts on the flow trace, expired VT > 6 mL/kg and < 10 mL/kg and satisfactory gas exchange (i.e. SaO2 > 90%, with pH > 7.35). The inspiratory trigger is set at the minimal level to avoid auto-triggering or that delivered by default by the ventilator, while the initial pressurization rate was adjusted according to subjective comfort. The fractional concentration of oxygen is such to achieve an SaO2 > 90%.

Standard Medical Therapy The usual standard of care. Sedatives and morphine could be administrated, but the dosage should be recorded in a data sheet (se data base)

The following variables will be recorded:

• Age, sex, chronic health care status measured using the McCrabe index, the number of co-morbid conditions, and the SAPS II index (Simplified Acute Physiology Score). Neurological status was assessed by the Kelly and Matthay scale, which is specifically designed for respiratory patients and Glasgow coma scale
• Causes of ARF (if apparent)
• Type and cancer status before respiratory failure
• Arterial blood was withdrawn from a radial artery at baseline and at fixed intervals (see protocol) and analyzed immediately (ABL 300 and ABL 625 Radiometer, Copenhagen, Denmark)
• Breathing frequency and hemodynamic variables (i.e. heart rate and blood pressure)
• Dyspnea score measured with a modified Borg scale
• Symptom Distress Scale (at enrollment and at discharge) as a surrogate of QoL "easy" measurements
• Palliative Prognostic Index (PPI)
• Types and dosages of sedatives or morphine
• Total days on NIV
• Total days of hospital stay
• Mortality at hospital discharge, at 6 and 12 months
• Causes of death ABG, respiratory rate and hemodynamic variables, Kelly score, Glasgow scale, PPI, Mc Cabe and dyspnea will be recorded at fixed intervals: baseline (T0), 1 h after beginning the treatment (T1), after 3 h (T2), after 24 h (T3), after 48 h (T4) and at discharge from the hospital (T5).

The SAPS II score is recorded at the end of the first day of admission. QoL questionnaire will be recorded at enrollment and hospital discharge Primary outcomes of the study were: hospital death, treatment failure, use of morphine or other sedatives and dyspnea and QoL scores. Treatment failure will be considered the "theoretical need of intubation"= MEET the INTUBATION CRITERIA defined as: 1) intolerance to NIV (only NIV group clearly) 2) no changes or worsening of ABG after 2 h of ventilation (pH < 7.35 or 20% decrease in PaO2/FiO2 ratio) 3) alteration of neurological status (1 point higher than admission in the Kelly score) 4) gasping for air or come. Secondary outcomes were: 6-month and 12-month mortality, the duration of hospital stay, changes of some physiological variables over time (i.e. ABG, respiratory rate) and QoL.

Inclusion Criteria:

Major criteria for enrollment into the study were one of the following:

• PaO2/FiO2 ratio < 250 + one of the two following:

  • dyspnea with recruitment of the accessory muscles and/or abdominal muscles recruitment; and
  • respiratory rate > 30 b/ min. Hypercapnia per se is not a criteria of inclusion, but it is not an exclusion criteria if chronic.

Exclusion Criteria:

• Potentially reversible causes of exacerbation such as (CPE, pneumonia or exacerbation of chronic pulmonary disorders) coma
• Refusal of treatment
• Inability to protect the airways
• An agitated or uncooperative patient
• Anatomical abnormalities interfering with mask fit
• Uncontrolled cardiac ischemia or arrhythmias
• Failure of more than two organs