Breathing patterns vary widely among normal subjects at rest, with Vt ranging from 5 to 17 ml/kg and RR ranging from 6 to 25 min-1 . Until the advent of PAV  it was not possible to determine the breathing pattern desired by ventilator dependent patients since, with conventional modes, breathing pattern is to a considerable extent determined by ventilator settings. When patients are placed on PAV at a high level of assist (i.e., in the absence of distress), the breathing pattern is found to be as variable among patients as it is among normal subjects. Data on a limited number of patients have been published [24, 58]. In a much larger group (n=80, unpublished observations) the range of patient-selected Vt was 0.23-0.891
(0.50±0.18 l), corresponding to 4-15 ml/kg (7.0±2.5 ml/kg). Undistressed RR ranged from 8.4 to 43.3 min-1 (23.8±7.5 min-1). In seven patients (9%) RRwas >35. Mechanical Ti, which during PAV is similar to nTI , ranged 0.48 to 1.61 sec (0.96±0.24 sec). Undistressed f/VT ratio ranged 10 to 171 (55±31). Eight patients (10%) had a ratio >110. Figure 2 shows two extremes of the range of breathing patterns observed. Undistressed Ve also varied widely (4.9 to 23.3,11.2±3.6 l/min).
1. The clinician setting the ventilator does not know what Vt the patient wants. He/she will tend to use a standard formula (e.g., 10 ml/kg). Given the very wide range of desired Vt (4-15 ml/kg), the percentage of patients who would be satisfied with a given fixed formula is directly related to the Vt of the formula. If clinician-chosen Vt is 15 ml/kg, all patients will receive their desired Vt. However, since virtually all patients would be satisfied with a smaller Vt, with this high Vt prescription most patients willbe subjected to unnecessary overdistension, with a consequent increase in risk of barotrauma. If the formula is 7 ml/kg, half the patients will get less than what they want and may become agitated. Of course, when a critically ill patient is agitated one does not know that this is because Vt is too small for the patient. The response would thus be to sedate the patient. The wide range of desired Vt, therefore, creates a no-win situation ifone uses afixedVT prescription. There is either unnecessary overdistension or unnecessary sedation, depending on the Vt of the prescription.
2. Delivering a Vt that is higher than that desired by the patient does not automatically result in a proportionate reduction in RR [5, 6, 12]. Much depends on ventilator Ti (see above). However, even if ventilator Ti is long (>2.0 sec), the reduction in RR is not enough to offset the higher Vt, so that PCO2 declines [5, 6]. It follows that providing a Vt that is higher than spontaneously chosen Vt will cause PCO2 to be lower than it would be otherwise. In the long term this may reduce the PCO2 set point. A decrease in PCO2 set point translates into a greater ventilatory demand at the time of weaning and this may affect weanabil-ity. Furthermore, the downregulation of inspiratory output will promote non-synchrony or recurrent central apneas in PSV (see above). In VCV, back-up rate will maintain ventilation but, in doing so, it will maintain and aggravate the relative hypocapnia, favoring a greater reduction in PCO2 set point in the long term.
3. The observation that a significant minority of patients (10%) choose a high RR and f/VT ratio without distress calls for caution when interpreting high RR and f/VT during a weaning trial. RR>35, or f/VT > 100, need not, per se, reflect weaning failure. We have seen many cases in which weaning failure, diagnosed on the basis of rate or f/VT criteria, turned out to be examples of spontaneously chosen breathing pattern; the same RR and f/VT were maintained as PAV assist was increased from 0 to near 100%. Such patients were successfully extubated notwithstanding the 'unfavorable' breathing pattern during the weaning trial.
4. The wide range of adequate Ve in different patients (as illustrated by the spontaneously chosen Ve) illustrates the difficulty that would be encountered with closed-loop ventilation systems that aim to target a set Ve:
Was this article helpful?