Introduction

Until fifteen years ago, mechanical ventilation was limited to the intensive care unit (ICU) because patients needed to be paralyzed and sedated. Over the last decade the great interest in non-invasive mechanical ventilation (NIV) has opened new horizons in the field of mechanical ventilation and the ways in which it can be applied. Indeed, as a result of various clinical and physiological evidence, NIV has become a first-line intervention in the management of severe exacerbations of chronic obstructive pulmonary disease (COPD) [1]. A recent International Consensus Conference in Intensive Care Medicine [2] concerning the use of noninvasive positive pressure ventilation (NPPV) in acute respiratory failure stated that "the addition of noninvasive positive pressure to standard medical treatment of patients with acute respiratory failure may prevent the need of intubation, and reduce the rate of complications and mortality in patients with hypercapnic respiratory failure". Moreover, it has been shown that NIV can be applied at an earlier stage in the evolution of ventilatory failure than would be usual when a patient is intubated and that the NIV canbe administered even outside the ICU. The literature about NIV in the treatment of 'pure hypoxic' acute respiratory failure is also abundant, but conclusive results have been reached only in acute respiratoryfailure due to exacerbation of COPD. A meta-analysis by Keenan and co-workers [3] showed that NIV can reduce the need for endotracheal intubation with survival benefits in this population. Unfortunately, a meta-analysis cannot focus on what is the ideal setting in which to apply NIV or the best population to receive it. Nonetheless, it should be emphasized that the positive results reported NPPV were observed only in selected COPD patients with the majority excluded. Patients with COPD needing immediate intubation have regularly been excluded from all studies published to date [4]. This means that wherever NIV is performed, facilities for prompt endotracheal intubation should be readily available. Indeed, major problems regarding the use of NIV out of an ICU are the possible need for quick intubation and ways to provide sufficient assistance during NIV. Absolute contraindications to NIV have been clearly recognized as:

• Cardiac or respiratory arrest

• Non respiratory organ failure (i.e., severe encephalopathy, severe gastrointestinal bleeding and hemodynamic instability with or without unstable cardiac angina)

• Facial surgery or trauma

• Upper airway obstruction

• Inability to protect the airway and/or high risk of aspiration

• Inability to clear secretions.

Apart from these contraindications, several studies have indirectly suggested or directly investigated the possible indicators of NIV success or failure. The recognition of these parameters maybe very important especially in environments outside the ICU, where it is mandatory to identify some 'quick' indicators of failure of NIV to avoid any undue delay in the time to intubation. In this chapter, we will discuss the utility of some predictors of NIV success or conversely of failure, depending on howyou approach the problem (i.e, the glass maybe half empty or half full), in both the situations of acute hypercapnic and 'pure hypoxemic' respiratory failure.

Acute Hypercapnic Respiratory Failure

In order to achieve the highest success rate during the application of NIV, an appropriate selection of patients who will benefit from this ventilatory technique must be made. The first step is to identify the patient in respiratory distress for whom ventilatory assistance is needed: clinical criteria include the presence of tachypnea, dyspnea, paradoxical abdominal breathing and accessory muscle activation while further indications can be given by the presence of respiratory acidosis (PaCO2 >55 mmHg and pH <7.35). The second step is the identification of patients needing immediate intubation for whom NIV is not the technique of choice (see the above listed contraindications to NIV).

The reported rates of NIV failure in most important clinical trials range from 5 to 40% [4-7], hence it is likely that the early recognition of a subset of'responders' could be helpful to avoid the delayed application of invasive ventilation in those patients who need endotracheal intubation.

Arterial blood gases

Respiratory acidosis is probably one of the most valuable indicators of the severity of COPD decompensation and all clinical studies report both pH and PaCO2 values sampled always atbaseline and aftera certain timelag (generallyone orafewhours) after the application of NIV. Ambrosino et al. [8] observed that, in a group of 47 patients with decompensated COPD, lower baseline values of pH and PaCO2 were predictive of NIV failure; in those patients who were more acidemic before starting ventilation (pH 7.22 versus 7.28) NIV subsequently failed. pH values recorded 1 hour after the initial trial of NIV also accurately identified those patients who could be successfully ventilated. Furthermore, using a logistic regression analysis, baseline pH and pH after 1 hr of NIV maintained a valuable predictive power with a high degree of sensitivity (87-93%, respectively) and specificity (54% and 82%).

Similar results have been obtained by Meduri et al. [9] in a group of 158 patients with acute respiratory failure sustained by different causes. In a subgroup of 74

patients with hypercapnic ventilatory failure, NIV failed in patients with higher baseline PaCO2 values and improvement of acidosis after a 2-hour NIV trial predicted a successful response to NIV.

Plant et al. [10] carried out a prospective multicenter randomized trial of NIV versus standard medical treatment on a population of 236 patients with an acute exacerbation of COPD and mild to moderate respiratory acidosis (pH range 7.25 7.35). They observed that a high degree of acidemia at study entry (pH<7.30) was associated with treatment failure and that improvement of pH after 4 hours of NIV was predictive of success. On the otherhand, in a multicenter epidemiologic survey, Carlucci et al. [7] showed, in a group of 108 patients treated with NIV, most of them affected by COPD, that changes in arterial blood gases after 1 day of NIV could not discriminate responders and non-responders (7.37 versus 7.34) but that pH on admission was significantly higher in favorable cases (7.36 versus 7.30).

In a small group of twelve decompensated COPD patients treated with NIV plus medical therapy, Soo Hoo et al. [11] had a 50% rate of successful cases and observed no difference in baseline pH and PaCO2 between failed and favorable cases. Yet, they were able to demonstrate that successful cases displayed a quicker correction ofacidosis.

Anton et al. [12] applied NIV to 36 hypercapnic COPD patients with a success rate of 77%; in order to find predictive criteria for NIV failure they devised a multiple regression model and concluded that improvement in PaCO2 and pH after one hour of NIV was a highly predictive index that enabled them to accurately predict NIV success in about 95% of cases.

While most authors have addressed the issue of'early' NIV failure, Moretti et al. [13] tried to analyze predictors of'late' failures, i.e., cases in which NIV failed a few days after its initial application, despite an initial improvement in clinical conditions and blood gas values. In a population of 134 patients with COPD exacerbation in whom NIMV was applied for > 24 hours, a subgroup of 31 patients did worse about 8 days after NIV application; a thorough evaluation of patient characteristics at study entry and at the moment of failure identified that a lower pH at admission, among with other variables, was predictive of late failure.

Table 1 summarizes the results regarding the achievement or not of statistical significance in predicting NIV success, for baseline pH and pH changes after a NIV trial.

Severity of Disease

The most commonly used indexes of severity of illness are APACHE II [14] and SAPS II [15] while other indexes, e.g., activities of daily living (ADL) [16] are specifically designed to indicate the degree of a patient's functional limitation. The association between indexes of severity of illness and NIV failure have been sought on the assumption that patients suffering from respiratory failure frequently have co-morbidities (malnutrition, cardiopathy, diabetes etc) and episodes of exacerbation are frequently associated with organ failures. Since during an episode of acute hypercapnic respiratory failure, the ventilator works mainly as an accessory muscle up to the resolution of the underlying decompensation, the greater the degree of

Table 1. The achievement (YES) or not (NOT) of statistical significance in predicting NIV success, for baseline pH or a change in pH during NIV

Baseline pH change after a NIV

pH trial

Table 1. The achievement (YES) or not (NOT) of statistical significance in predicting NIV success, for baseline pH or a change in pH during NIV

Baseline pH change after a NIV

pH trial

Soo Hoo [11]

No

Yes

Moretti [13]

Yes

No

Anton [12]

No

Yes

Ambrosino [8]

Yes

Yes

Plant [10]

Yes

Yes

Carlucci [7]

Yes

No

Meduri [9]

Yes

Yes

organ failure, the slower will be the recovery process and the less likely will be the success of NIV.

A positive correlation between severity of underlying disease and NIV failure has been demonstrated by some authors. Moretti et al. [13] found that the presence of one or more complications at admission (i.e., severe hyperglycemia) together with marked functional disability (low ADL scale) were strong predictors of late NIV failure. A significantly greater severity of illness among patients who failed to improve with NIV was found by Ambrosino et al. [8] (mean APACHE II 24 versus 18 in successful cases) and by Soo Hoo et al. [11] who reported a mean APACHE II score of 21 in NIV failures and of 15 in successful cases. In the previously quoted epidemiologic survey [7], Carlucci et al., applying a multiple regression analysis, showed that SAPS II was an independent predictive factor of NIV failure. In another randomized prospective study in a general ICU, Conti et al. [17] treated 49 decompensated COPD patients (mean pH 7.2) with invasive (26 pts) or non-invasive (23 pts) ventilation after failure of medical treatment performed in the emergency department. The failure rate in the NIV group was 52% and patients who needed endotracheal intubation after a trial of NIV had a significantly higher SAPS II score (mean value=39) compared to patients who avoided intubation (mean value=35). However, using APACHE II, Anton et al. [12] and Meduri et al. [9] did not find any correlation between severity score andNIV failure. Similarly, Benhamou et al. [18], using SAPS II, was unable to find a link between patients' clinical status and NIV failure.

Table 2 summarizes the results concerning the achievement or not of statistical significance in predicting NIV success, for various indices of disease's severity.

Table 2. The achievement (YES) or not (NOT) of statistical significance in predicting NIV success, for various indices of disease's severity. ADL: activities of daily living scale

Was this article helpful?

0 0

Post a comment