Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

Highlighted lines and questions below provide links to the pertinent description of criteria in The EBM User's Guide, now available at the Canadian Centres for Health Evidence

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

Noninvasive ventilation in severe hypoxemic respiratory failure: a randomized clinical trial.

Ferrer M, Esquinas A, Leon M, Gonzalez G, Alarcon A, Torres A.

Am J Respir Crit Care Med. 2003 Dec 15;168(12):1438-44Chest 2002;121: 858-862 [abstract]

Reviewed by Hanna Sahhar MD, Advocate Hope Children's Hospital, Chicago, IL

Review posted February 14, 2004

---

I. What is being studied?:

The study objective:

To assess the efficacy of noninvasive ventilation (NIV) in preventing intubation as the primary end-point variable compared with a standard regime consisting of high-concentration oxygen therapy.

The study design:

A prospective, randomized controlled study was conducted in three adult ICUs.

The patients included:

Patients (n=105) with severe Acute Hypoxemic Respiratory Failure (AHRF), which is defined as PaO2 less than 60 mm Hg for more than 6-8 hours or arterial oxygen saturation by pulse oximetry persistently less than 90% while breathing conventional Venturi oxygen at a maximal concentration of 50%.

The patients excluded:

Exclusion criteria were (1) hypercapnia (PcO2 more than 45 mm Hg) on admission; (2) need for emergency intubation; (3) recent esophageal, facial, or cranial trauma or surgery; (4) severely decreased consciousness (a Glasgow coma scale of 11 or less); (5) severe hemodynamic instability despite fluid repletion and use of vasoactive agents; (6) a lack of cooperation; (7) tracheotomy or other upper airway disorders; (8) severe ventricular arrhythmia or myocardial ischemia; (9) active upper gastrointestinal bleeding; (10) an inability to clear respiratory secretions; and (11) more than one organ dysfunction in addition to respiratory failure.

The interventions compared:

Patients were randomly allocated within 24 hours of fulfilling inclusion criteria either to the noninvasive ventilation group or the control group. In the NIV group, patients were ventilated using the bilevel positive airway pressure mode (BiPAP) by face or nasal mask. The control group received oxygen using high concentration sources while spontaneously breathing.

The outcomes evaluated:

The main outcome variables that the investigators evaluated were (1) the intubation rate and indication; (2) the duration of ICU and hospital stay; (3) the ICU mortality and 90-days survival rate; (4) the complications after patients were entered into the study, in particular the incidence of septic shock; and (5) the clinical improvement.

II. Are the results of the study valid?

Primary questions:

1. Was the assignment of patients to treatments randomized?

Yes. The investigators stated that the patients were randomly allocated and assigned to one of the two groups. They did not specify the randomization method (computerized or not).

2. Were all patients who entered the trial properly accounted for and attributed at its conclusion?

Was followup complete?

Yes. All patients in both groups were followed up until 90 days after entry into the protocol or death.

In the NIV group, 51 patients entered the study; 10 died and 41 completed the 90-days survival, while in the control group, 54 patients entered the study, 21 died and 33 completed the 90-days survival evaluation.

Were patients analyzed in the groups to which they were randomized?

Yes. Even though, two patients from the control group used NIV because of spontaneous breathing failure and were spared intubation, and one patient from the NIV group did not tolerate NIV before achieving criteria for clinical improvement (SpO2 of more than 92% or PaO2 of more than 65 mm Hg while spontaneously breathing) without intubation, all three patients kept in their own randomized group for analysis.

Secondary questions:

3. Were patients, health workers, and study personnel "blind" to treatment?

No. The study was an open clinical trial, which might have led to possible bias. Despite the fact that the investigators predefined the criteria for all relevant interventions, clinical decisions, and outcome variables, this bias could not be entirely controlled.

4. Were the groups similar at the start of the trial?

Yes. No significant differences between the two groups were shown in age, sex, severity of illness, presence of pulmonary infiltrate, underlying diseases, respiratory frequency, heart rate, blood pressure, arterial blood gases, and causes of Acute Hypoxemic Respiratory Failure.

5. Aside from the experimental intervention, were the groups treated equally?

No specific guidelines or protocols for treatment were predefined. The investigators mentioned in the introduction of the article that concomitant treatments for the underlying disease were instituted, but cointerventions were not clearly defined. Reporting concurrent therapies is crucial for a non-blinded trial to reduce the risk of bias.

III. What were the results?

1. How large was the treatment effect?

The primary end-point variable for this study is the decrease in the intubation rate. Patients in the noninvasive ventilation (NIV) group had an intubation rate of 25% versus 52% in the control group (p = 0.010). The relative risk reduction (RRR) of using NIV is 51%, the absolute risk reduction (ARR) is 26% (95% CI 8%, 44%), and the number needed to treat (NNT) is 4 (95% CI 2, 12). The multivariate analysis that the investigators used showed that NIV was independently associated with decreased risk of intubation; adjusted odd ratio 0.20 (95% CI 0.07, 0.58; p value = 0.003).

The mortality rate was also lower in the NIV group (18%) versus 39% in the control group. We found that the RRR is 55%, the ARR 21% (95% CI 5%, 38%), and the NNT is 5 (95% CI 3, 22). By using the log-rank test, the investigators found that the cumulative survival probability within 90 days after entry into the protocol was significantly higher in the NIV group (p = 0.025).

Another important outcome variable they looked at is the incidence of septic shock. The NIV group had a 6% incidence of septic shock versus 31% in the control group (p = 0.028). Using NIV results is 63% fewer episodes of septic shock (RRR) than the conventional oxygen regime (control). The ARR is 19.7% (95% CI 4.5%, 34.9%). The NNT is 5 (95% CI 3, 22).

The investigators avoided the use of Pao2/FiO2 ratio as an inclusion criteria arguing that P/F ratio is dependent on the Fi02 delivered and may not always reflect life-threatening consequences of the deterioration of oxygenation, but they used the P/F ratio as an indicator for improvement in both groups and that there was a higher improvement in the NIV group in the first 72 hours and even significantly higher in the first 3-4 hours after randomization (p = 0.029). After Bonferroni correction, P/F ratio remained significantly higher at 6-8 and 24 hours after randomization in the NIV comparing to the control group. A similar result in the improvement of the respiratory rate was also evident.

There was no decrease in the ICU and in-hospital stays between the two groups in the overall population, but in-hospital stay among ICU survivors decreased in the NIV group (p = 0.043).

2. How precise was the estimate of the treatment effect?

There was a difference in the primary end-point between the two groups. Although the lower boundary for the 95% confidence interval for the ARR is just 8%, it is still clinically significant. It is noteworthy to mention that the study was powered to analyze the overall population; because of the relative heterogeneity of patients with AHRF, the small sample sizes of the subgroups were too small to be analyzed conclusively.

IV. Will the results help me in caring for my patients?

1. Can the results be applied to my patient care?

When trying to apply the results of this study to our Pediatric population, many factors should be taken into consideration. The disease spectrum in adulthood is different than that in childhood, even though the pathophysiology may be similar. ARDS, as a common cause of respiratory failure in our PICU, is a consequence of many etiologies; e.g., trauma, poisoning, infection. Our clinical experience of early intubation, even using high frequency ventilation, seems to yield a better outcome in this subset of patients, but we are still waiting for prospective clinical studies and subsequent statistical evidence to support our observation and practice.

Facemasks are often less well tolerated in children - in particular in infants - than adults. In addition, children may develop gastric distension to a certain degree from non-invasive ventilation.

In our unit, there is an increase in the trend toward the use of the noninvasive ventilation in some cases of status asthmaticus, acute chest syndrome, and pulmonary edema to prevent intubation and its complications. This subset of patients has intact respiratory drive, and the respiratory failure is due to increased work of breathing and ultimately mechanical failure.

2. Were all clinically important outcomes considered?

All important outcomes were considered. The main complication of the noninvasive ventilation was the nasal bridge injury in 25% of the patient. No information was available for follow up of this complication. Gastric distension and conjunctivitis seemed trivial and transient complications of NIV in the adult population. In pediatrics, gastric distension could be a significant complication that can lead to vomiting and possible aspiration. In addition, the mechanical obstacle for lung inflation against the distended stomach may increase the work of breathing and exaggerate the pre-existing disease process.

3. Are the likely treatment benefits worth the potential harms and costs?

When weighing the benefits versus the risks, we found that the risks of NIV were mild to moderate nasal bridge injury (25% of patients), conjunctivitis (6%), and gastric distension (2%). The benefits of the decrease in intubation rate, septic shock and mortality rate outweigh the transient risks.

The authors found that the in-hospital stay among ICU survivors decreased in the NIV group (21.1 ± 14.8 days in the NIV group and 30.2 ± 21.3 days in the control group) (p = 0.043). The cost of using BiPAP machine, even though it is not mentioned in the article, is likely to represent a cost savings compared to additional days in the hospital.

We are and we will keep using NIV in our pediatric intensive care unit whenever it seems appropriate to try and is feasible. This study lends indirect evidence that such practice is warranted until pediatric studies are performed.

Comments

Back to the EB Journal Club Index