Nava, S., Ferrer, M., Esquinas, A., Scala, R., Groff, P., Cosentini, R., . . . Grassi, M. (2013). Palliative use of non-invasive ventilation in end-of-life patients with solid tumours: A randomised feasibility trial. The Lancet Oncology, 14, 219–227. 

To determine the acceptability of solely using palliative noninvasive ventilation (NIV) versus oxygen therapy to manage breathlessness in patients with end-stage cancer and its effects in reducing dyspnea and opioid requirement

• All patients received a 5–10 minute demonstration on NIV prior to randomization.
• Patients were randomly assigned to either NIV (pressure-support mode and initiation based on patients’ request and mask comfort) or oxygen (Venturi or reservoir mask) group to achieve oxygen saturation > 90%.
• Based on a computerized randomization sequence, patients were further stratified and randomly assigned to hypercapnic (PaCO2 > 45 mm Hg) or nonhypercapnic (PaCO2 < 45 mm Hg) groups.
• An independent biostatistitian placed patients’ information in an opaque, sealed envelope.
• Subcutaneous morphine was given to reduce dyspnea scores by at least one point on the Borg scale (BS), and the total dose requirements over the first 48 hours were calculated.
• Arterial blood gas, vital signs, quantity of secretion, and dyspnea scores were measured at baseline and at one hour, 24 hours, and 48 hours.
• Mortality causes were recorded in the hospital and at three and six months after discharge.

• N = 200
• MEAN AGE = 71 years (NIV group), 70 years (oxygen group)
• MALES: 62%, FEMALES: 38%
• KEY DISEASE CHARACTERISTICS: Solid tumor cancers included lung, gastrointestinal, breast, head and neck, and esophageal. Respiratory failures included obstructive bronchus, carcinomatous, lymphangitis, and pleural effusion. Life expectancy was less than six months.
• OTHER KEY SAMPLE CHARACTERISTICS: Inclusion criteria were a P/F ratio of 1:250 and one of the following: dyspnea with a BS ≥ 4, signs of respiratory distress, or a respiratory rate ≥ 30. Exclusion criteria were reversible respiratory failure such as cardiogenic pulmonary edema or the exacerbation of chronic pulmonary disorders, treatment refusal, a weak cough reflex, agitation or uncooperative behavior, anatomical abnormalities with mask fitting, uncontrolled cardiac ischemia or arrhythmias, ≥ 2 organ failure, opioid use within past two weeks, adverse effects to opioids, substance misuse history, contraindication to morphine use, acute renal failure, and recent head injury.

• SITE: Multi-site  
• SETTING TYPE: Inpatient  
• LOCATION: Seven intensive care units and critical care units in Italy, Spain, and Taiwan

• PHASE OF CARE: End-of-life care
• APPLICATIONS: Palliative care

Multi-center, randomized, controlled trial

• McCabe Cyclomatic Complexity Index (MCCI) to record demographic information
• Simplified Acute Physiology Score II (SAPS) and Palliative Prognostic Index Score (PPIS) to determine mortality risk
• Kelly and Matthay Scale (KMS) to assess neurological status in patient with respiratory disease
• Modified Borg Scale (BS) to assess intensity of dyspnea
• Symptom Distress Scale (SDS) self-administered questionnaire to assess cancer-related symptoms

• In the NIV group, 11 patients (11%) dropped out of the study, but there was no attrition in the oxygen group.
• The NIV group had a statistically significant reduction in dyspnea levels at one hour, 24 hours, and 48 hours (BS = -0.58; 95% CI -0.92, -0.23; p = 0.0012). The greatest reduction was seen in the hypercapnic NIV group after the first hour (BS = -0.91; 95% CI -1.42, -0.46; p < 0.0001). 
• The total amount of morphine required over 48 hours was lower in the NIV group (SD = 37.3 mg) compared to the oxygen group (SD = 67.1mg) with a mean difference of -32.4 mg (95% CI -47.5, -17.4).
• In-hospital mortality was the same between the groups.
• The NIV hypercapnic group had better expected survival rates than the oxygen hypercapnic group.
• At discharge, symptom distress scores were statistically improved in the NIV group compared to the oxygen group (SD = 10 versus SD = 7.6; p = 0.001).

NIV was feasible and effective in decreasing dyspnea intensity and reducing morphine requirements in patients with end-stage cancer experiencing respiratory failure. However, additional studies validating these findings and determining the effects of NIV on survival and quality of life are needed.

• Risk of bias (no blinding)
• Key sample group differences that could influence results
• Findings not generalizable
• Intervention expensive, impractical, or training needs
• Subject withdrawals ≥ 10% 
• Other limitations/explanation: The benefit of performing a blinded study using a sham ventilation could have potentially caused harm to the patients in the oxygen group. Therefore, the researchers chose to maintain patient safety.
• The NIV group had an attrition bias. Also, the duration and timing of the NIV treatment was based on patients’ preferences and mask comfort, leading to a threat in internal validity. The NIV treatment may not be practical or cost effective in all palliative care centers because specialized training and equipment are needed. 

This study offers clinicians a treatment modality that can be used in adjunct with opioids to significantly reduce breathlessness in patients with end-stage cancer. Additional studies are needed to determine the specific patient population that would benefit the most from this treatment, its cost effectiveness, patient satisfaction, the adverse effects of NIV, and the survival rate.