INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

It has previously been demonstrated that in normal subjects, mechanically ventilated on a volume-cycled ventilator, increasing inspiratory flow rate has an excitatory effect on respiratory rate (1). The current study was undertaken to answer two questions regarding the effect of inspiratory flow rate on respiratory rate: (1) Is this response present not only in normal subjects but also in patients with respiratory disease? (2) Is this response mediated by mechanisms that are independent from effects of flow rate on upper airway receptors?

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Eight patients were studied (Table 1). All were patients in the intensive care unit who were undergoing mechanical ventilation for respiratory failure. All were orotracheally intubated. Seven patients were studied while ventilated on a Bear-2 ventilator, while one patient was ventilated with an Infrasonics, Inc., Adult Star ventilator. Informed consent to perform the study was obtained from either patients or their families. As a result, most patients were aware that changes would be made to the flow rate of the ventilator. They did not know when these changes would be made, nor were they aware of the specific responses that were being measured.

All patients were changed to volume-cycled ventilation in the assist-control mode. The humidifier was disconnected from the ventilator circuit. Back-up rate was set at 0.5 breaths/min, ensuring that all breaths would be patient-triggered. Initial flow rate was set at 60 L/ min. Tidal volume (VT) was initially set at 7 cc/kg. For patients who were alert enough to communicate, VT was then adjusted up or down until the patient was comfortable. Once at VT was chosen, it was kept constant throughout changes in flow rate. Because changes in flow rate would result in different levels of airway pressure (Paw), the amount of VT delivered to the patient might vary as a result of different amounts of gas being lost to tubing distention in the ventilator circuit at different Paw. In an attempt to minimize these differences in VT, short lengths of low-compliance ventilator tubing were used. Positive end-expiratory pressure (PEEP) was maintained at the prestudy level (5 cm H₂O in seven patients and 0 cm H₂O in one). The FI_O2 was set at a level that maintained oxygen saturation about 94% and, once set, was kept constant for the duration of the experiment. Ventilator trigger sensitivity was set at maximum, which resulted in triggering when Paw dropped below 1 cm H₂O. All patients were ventilated at these initial settings for approximately 20 min before any changes in flow rate were made. All flows were delivered with a square-wave flow-time profile. Flow was monitored between the Y-connector and the endotracheal tube with a heated pneumotachograph (Hans-Rudolph 3700, Kansas City, MO). Both inspiratory and expiratory flows were monitored. Volume was derived from the electronic integration of flow. The Paw was monitored at the level of the pneumotachograph. All signals were sampled at 50 Hz (CODAS DATAQ Instruments, Inc., Akron, OH) and stored on a computer disc for later analysis.

Once patients were acclimatized to the new ventilator settings, trials were undertaken in which the flow rate was changed from a baseline flow of 60 L/min. Trials involved both increasing flow to 90 L/min and decreasing it to 30 L/min. Flow was maintained at 30 or 90 L/min for a minimum period of 3 to 4 breaths, and then returned to 60 LPM for at least 60 s prior to the next flow change being attempted. We obtained a minimum of five flow rate transitions at both 90 and 30 L/min in each patient. In three patients who appeared comfortable with flow rates of 30 L/min for an extended period of time, transitions were also undertaken from a baseline flow of 30 L/min to a flow of 90 L/min, again for a duration of 3 breaths.

For each trial involving a transition in flow rate, 6 breaths were analyzed, 3 breaths at the baseline flow rate and 3 breaths at the new flow rate. Analysis included a breath-by-breath determination of Ttot, mechanical TI (onset of inspiratory flow to onset of expiratory flow), mechanical TE (balance of Ttot), inspiratory VT, and expiratory VT. Trials in which there were inspiratory efforts that failed to trigger the ventilator (transient decrease in Paw or expiratory flow with no triggering) were discarded. For each trial, we calculated mean values for the 3 breaths at each of the two different flow rates. These mean values from similar trials (60 to 30, 60 to 90, and 30 to 90 L/min) were pooled and mean values calculated for each patient. Each patient's responses were assessed for statistical significance by the paired t test. Group responses for each flow transition were assessed also by the paired t test (n = 8 for 60 to 30 L/min and 60 to 90 L/min, n = 3 for 30 to 90 L/min).

To minimize distractions, nursing interventions such as suctioning and turning the patient were carried out before data collection. The state of wakefulness or sleep was ascertained by behavioral criteria; flow transitions were carried out only when patients were awake.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The total number of trials undertaken varied among patients, ranging from 9 trials in patient 5 to 45 trials in patient 8. The mean number of trials per patient was 11.2 for the 60 to 30 L/ min transitions, 16.2 for the 60 to 90 L/min transition, and 11.7 for the 30 to 90 L/min transition. Figure 1 shows sample graphs from flow transitions in one patient.

View larger version (51K): [in this window] [in a new window]   Figure 1.   Examples of response to changes in inspiratory flow in patient 7. (A) Flow rate was changed from 60 to 30 L/min. (B) Flow rate was changed from 60 to 90 L/min. Note that respiratory rate changes in the same direction as the change in flow. Arrow indicates flow transition. The scale underneath B indicates time in seconds.

In all eight patients, respiratory rate changed in the same direction as inspiratory flow rate (Figure 2, Table 2). For the patients as a group, respiratory rate at an inspiratory flow of 30 L/min was significantly less than it was at 60 L/min (p < 0.001), with a mean decrease in rate of 3.4 breaths/min (range, 1.8-6.6). The decrease in respiratory rate with this flow transition was significant in all eight patients individually as well (p < 0.05). For the eight patients as a group, respiratory rate at a flow rate of 90 L/min was significantly greater than at 60 L/min (p = 0.03), with a mean increase of 2.3 breaths/min (range, 0.52-8.3). Six of the eight patients had statistically significant (p < 0.05) increases in respiratory rate individually with this flow transition. In the three patients whose respiratory rates were compared between a baseline flow of 30 L/min and a flow rate of 90 L/min, there was a marked increase in rate, with a mean increase of 5.5 breaths/min. The increase in respiratory rate in these three patients when flow was changed from 30 to 90 L/min was slightly greater in magnitude than the sum of the changes in rate between 30 and 60 L/min and between 60 and 90 L/min. This change in rate did not achieve statistical significance (p = 0.06) for the group, due to the low number of patients, but it was statistically significant (p < 0.05) for each of the three patients individually (Figure 2).

View larger version (19K): [in this window] [in a new window]   Figure 2.   Individual responses of respiratory rate to the three flow transitions. An * over a line denotes significant change (p < 0.05).

By design, there was a significant decrease in TI with increasing flow rates (Table 2). Figure 3 compares the actual change in TE (black bars) with the change predicted had there been no change in respiratory rate (grey bars; expected Te = Ti). The TE response was not consistent. In some subjects the change was minimal, while in others it was opposite in direction to what was expected. Even when TE changed in the expected direction, the response was much smaller than expected. The group TE responses were not significant with any flow transition.

View larger version (37K): [in this window] [in a new window]   Figure 3.   Predicted change in TE (grey bars) versus actual change in TE (black bars) for individual patients with the three flow transitions. The predicted values are based on an unchanged Ttot; they equal TI. Note the lack of a consistent relation between predicted and observed values.

Inspiratory and expiratory tidal volumes remained fairly constant at the three different flow rates (Table 2). As expected, peak Paw changed in the same direction as flow rate.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our results indicate that, for intubated patients with respiratory disease who are ventilated with volume-cycled ventilators and who are triggering the ventilator, increases in inspiratory airflow cause an increase in respiratory frequency. The response is observed within the first breath and is maintained for a least 3 breaths.

Puddy and Younes studied the response to changes in inspiratory flow rate in humans and demonstrated that in normal subjects, mechanically ventilated on a volume-cycled ventilator, increases in flow rate had an excitatory effect on respiratory frequency (1). The response was largely complete within 1 breath, suggesting a neural origin for the effect. The increase in respiratory rate resulted primarily from a decrease in TI, although TE did decrease somewhat as well. More recently, it has been shown that this response is not affected by breathing route (nose versus mouth), temperature of inspired gas, VT at which the response is studied, or upper and lower airway anesthesia (2). The effect has been shown to be attenuated but present in non-REM sleep (3). This latter finding confirms that the response is reflexic in origin.

Our study extends the previous observations by demonstrating that the response is not related to a reflex mediated by upper airway receptors, since it is present in intubated patients. Although an earlier study (2) showed preservation of the reflex after surface anesthesia of the airway, the possibility remained that it was mediated by deeper airway receptors not accessible to surface anesthesia. The fact that this response is preserved in intubated patients is not unexpected. There is evidence that airflow through the upper airway has an inhibitory effect on respiratory rate (4). Therefore, shielding the upper airway from the inspiratory flow of air would, if anything, be expected to enhance the response, not attenuate it.

In comparing the changes in respiratory rate that we observed in intubated patients with the rate changes observed in normal subjects (1), the mean change in rate was slightly greater at the 60 to 30 L/min transition in our patients (3.4 versus 2.1 breaths/min), but was slightly less at the 60 to 90 L/min transition (2.3 versus 3.2 breaths/min).

Could the effect we demonstrated be due to factors other than changes in flow rate? There is little evidence to suggest this. Tidal volumes were very similar at the different flow rates. Chemoreceptor influences on rate were likely not a factor either, since the change in flow rate lasted only 3 breaths, making changes in CO₂ or O₂ an unlikely mediator of rate changes. Attempts were made to minimize external stimuli that might result in behavioral stimulation of respiratory rate, making it unlikely that this had a significant impact on our results. In any event, this effect would likely be random and therefore not have any appreciable effect on the multiple trials undertaken in each of the eight patients.

The presence of this response in ventilated patients has some practical implications:

1. Inspiratory flow is often increased to provide more expiratory time. Our data demonstrate that this practice actually has a variable effect on TE (Table 2, Figure 3). This effect is due to two appositional tendencies. On the one hand, increased flow rates result in a reduced mechanical TI which, in theory, allows more time for expiration prior to the next breath. On the other hand, increased flow rates tend to increase respiratory rate. Whether TE increases or not depends on which of these two tendencies prevails, which in turn depends on how strong the excitatory response to higher flow rates is in any individual patient. It is important to note that even in those patients in whom TE increased, the increase was far less than one would expect in the absence of an increase in respiratory rate.
2. If respiratory rate increases in response to increased flow rates, minute ventilation will be increased due to the fact that VT is held constant on volume-cycled ventilators. This will, over time, result in a fall in PCO₂, provided that dead space and CO₂ production remain constant. The resulting hypocapnia could be a source of weaning difficulties (5).

An important clinical question is whether changes in respiratory rate are sustained or attenuated over time due to reciprocal changes in PCO₂. We did not undertake to study rate changes for extended periods of time due to the practical difficulties in the intensive care unit of preventing changes in level of wakefulness and environmental stimuli. However, we feel such an adaptation to the response is unlikely for two reasons. First, the influence of PCO₂ on rate in ventilated patients and normal subjects has been shown to be minimal in the hypocapnic and eucapnic range (6, 7). Second, no adaptation to the response was seen in normal subjects after 3 to 4 min (1).