Pressure support is designed as a true assist form of

Transcription

1 Effects of nitial Flow Rate and Breath Termination Criteria on Pressure Support Ventilation* Neil R. Macintyre, M.D., EC.C.P.; and Li-ng Ho, M.D.t To assess whether adjustments in the initial How rate or breath termination criteria affected patient-ventilator synchrony, we studied the ventilatory pattern response to PS in 33 patients under two sets of circumstances:during seven different levels of delivered initial PS How and during PS termination at 50 percent and at 25 percent of peak How. n the study on initial PS How, we found the following: (a) an optimal initial PS How could be defined for a given level of PS that resulted in the patient obtaining maximal pressure and volume from the ventilator; (b) initial PS Hows above and below this optimal How were associated with faster breathing frequencies, shorter inspiratory times, smaller tidal volumes and a tendency for airway pressure to not reach the selected PS level; and (c) optimal initial PS How was fastest in patients with the lowest compliancesand the most active ventilatory drives. Changing PS termination criteria from 50 to 25 percent of peak How had minimal effectson the ventilatory pattern or synchrony. Weconclude that the initial PS How to achieve the selected PS level is important in patient-ventilator synchrony but that termination criteria set between 25 and 50 percent of peak How is not. (Chest 1991; 99:134-38) PS = pressure support; MV= minute ventilation; f = respiratory frequency Pressure support is designed as a true assist form of mechanical ventilation.!" n concept, PS "boosts" a patients ventilatory effort through the application of a clinician set level of inspiratory airway pressure. With PS, the patient maintains control of inspiratory and expiratory timing while the ventilator adjusts delivered flow to patient demand in order to maintain a constant inspiratory airway pressure. A potential advantage of pressure assisting with PS as compared to volume assisting with conventional volume cycled ventilation is that the delivered How pattern with PS might synchronize better with a patients ventilatory drive.5.6this concept is based on the observation that once a patients ventilatory effort is initiated, inspiratory muscle contraction continues regardless of the presence of a ventilator-assisted breath."?" t then follows that a ventilator-assisted breath, which sets pressure as the independent variable (ie, PS) and lets flow and volume be dependent on this ventilatory effort, might synchronize with this effort more readily than a ventilator assisted breath which sets flow and volume as independent variables unaffected by patient effort (ie, a volume-assisted breath). Although this synchrony concept is attractive, two features of current PS design may limit its application *From the Department of Medicine and Respiratory Care Services. Duke University Medical Center, Durham, NC. tpresently at Section of Respiratory Therapy. Veterans General Hospital. Taipei, Taiwan. Manuscript received January 22; revision accepted July 19 Reprint requests: Dr. Mcintyre. Box 3913 Duke University Medical Center, Durham (Fig 1). The first is that the flow to initially reach the selected inspiratory pressure is fixed, usually at the ventilator maximum. fthis initial flow is below patient demand, we might speculate that the observed airway pressure rise will be slow and may not even reach the selected pressure level before the patient terminates the breath. On the other hand, ifthe initial flow rate is in excess of patient demand, we might speculate that such rapid airway pressurization could disrupt the inspiratory muscle contraction pattern and result in a premature termination of inspiratory effort. Alternatively, an initial PS flow well above patient demand could also cause premature breath termination because flow termination criteria would also be inappropriately high. Under any of these circumstances, the ability of a PS breath to provide a desired level of muscle "unloading" would be compromised.v" The second potentially limiting design feature is that PS termination criterion is fixed and usually arbitrarily set at 25 percent of peak flow (ie, the ventilator is flow cycled 08). 2 n some patients, this may be too early and could result in additional inspiratory muscle load ifthe inspiratory effort persists after the ventilator has cycled off. Conversely, in other patients, this may be too late and could result in unnecessary expiratory work ifexhalation is attempted while PS is still being delivered. n order to study these patient-ventilator interactions during application of PS, we measured airway pressure and the ventilatory pattern response of 33 patients to adjustments in the initial inspiratory flow or to adjustments in breath termination criteria. 134 nitialflowrate,breath Termination Criteria,and Pnlssu8 Support Ventilation (Macintyre. Ho)

2 Pressure Flow ns Respome 7bChangesin the nitial FlowRate Seventeen patients were ventilated with the RSA ventilator. This ventilator has an adjustable initial inspiratory Bow during PS that ranges from maximal delivery capacity to 25 to 30 percent of this value. The selected pressure support level (PSSEL) was that which provided a VT of 9 to 12 mllkg at the maximal Bow setting.,,, \" : Tme ~ Time Exp., FCURE 1. Schematic depicts the effects of varying inspiratory flow andbreath termination criteria on a pressure-supported spontaneous breath. n the top ponel, airway pressure is depicted; in the bottomponel, flow is depicted over single breath. lbint A indicates the onset of a spontaneous inspiratory effort and the initiation of pressure support. lbint B indicates attainment of the selected pressure support level. The solid line indicates rapid initial ventilator flow delivery and a rapid attainment of the pressure support level; thejmhed line indicatesslower nitial ventilatorflowdelivery and a slower attainment of the pressure support level. Note that after the selected pressure support level is reached, this airway pressure plateau is maintained through ventilator adjustments in inspiratory flow. lbint C indicates pressure support termination. The Olid line rebects termination at a lower fraction of peak Bow than the dottedline. PATENTS AND METHODS Thirty-three clinically stable, mechanically ventilated patients were studied. nitial data were collected while the patient was receiving hislber baseline mechanical ventilatory support (all patients were either on high levels of PS or on synchronized intermittentmandatoryventilationwith lower levels ofps). Baseline dataincluded: diagnosis, total minute ventilation determined by the ventilators exhaled volume monitor (MV), static respiratory system compliance (Cas, volume controlled tidal volume of 10 to 12 mlllcg! plateau airway pressure), and arterial blood gas values (Po., Pco., ph). During all studies, positive end-expiratory pressure (PEEP) and Flo. remained constant. Standard circuit tubing with a heatmoisture exchange humidifier was used in all patients. After a period of stabilization, the initial PS flow rate was then decreased in six discrete steps until minimal flow was provided. The PSSEL was not changed during these reductions. At each Bow setting, the ventilatory pattern was allowed to stabilize and then measurements were made of patients VT, f, and airway pressure. Using the airway pressure tracing, we calculated the inspiratory PS triggering pressure (PrJuc), the inspiratory time ('i) and actual airway PS level attained (1'Ps). At each setting we also estimated ventilator "work" or "output" by the measured pressure volume product (PV='PsX VT). The optimal initial PS 80w setting was defined as that setting resulting in the maximal PY. The VT, f, and 11 with the optimal initial PS Bow setting were compared to those at the maximal and minimal flow settings by paired Student's Hests using the Bonferroni adjustment for significance levels when using paired data twice (ie, p<o.025 was considered significant). This adjustment protects the overall significance level at.05 for both tests taken simultaneously. n addition, theoptimalinitial PS Bowsettingwas comparedby linearregression analysis to indices of ventilation load (ie, baseline MV and Cas) and ventilatorydrive (PTJuc). A p value <0.05was considered significant for this analysis. Peale ventilator delivered BowMat each setting was measured by a differential pressure Bow transducer in 11 patients. Respome 7bChanges in the Breath Termination Criteria Sixteen patients were ventilated with the Bear3 which had been specially modified with an adjustable PS termination control at either 25 or 50 percent of the peak Bow. The selected pressure support level (PSSEL) was that which provideda VTof9 to 12 mllkg with the PS termination control set at 25 percent peak 80w. Patients were then ventilated 20 to 30 minutes under these conditions and then 20 to 30 minutes with the PS termination control set at 50 percent peale 8ow. Measurements of VT, f, and airway pressure were recorded as in the initial PS flow study at the end of each ventilation period. Paired data were compared by Student's e-test. A p value <0.05 was considered significant. RESULTS Baseline characteristics ofthe 33 study patients are given in Table 1. \fzrying nitial PS Flow Rates The range of initial PS How delivered by this ventilator during these experiments was 34 to 104 U Table -Clinical Characteriatica at Ba8eUne of Study popu/ationa (DGta are Numbera or Mean ± SV) nitial PS Flow Study PS Breath Termination Study Number Diagnoses Pneumonia 1 4 Adult respiratory 7 5 distress syndrome Cardiac edema 3 2 Central nervous 4 3 system dysfunction Other 2 2 MY, Umin 13.3± ±5.8 Cas, mllem H.O 41.2± ± 10.1 PaO/Fo., mm Hg 281.3± ± PSSEL, em H.O 22.8± ±7.6 CHEST JANUARY,

3 #3 VT =930 #7 VT=620 #6 VT =910 #7 VT=900 /ld\, o 5 10 Time (sec) A./\ o 5 10 FGURE 2. Examples of airway pressure tracings in two patients with different initial PS flows during PS ventilatory support. nitial PS flows are expressed as one of seven discrete settings ranging from maximal (#1) to minimal (#7). For these two patients, a maximal, optimal and minimal initial PS How setting is depicted. n the top panels, the optimal initial PS How setting for this patient was the third fastest (#3). Note that the attained Prs and VT appeared highest, and the Ti longest at the optimal initial PS Howsetting (Fig 3 and 4). n the bottom panels, the optimal initial PS How setting for this patient was the second slowest (#6). Again note that the attained Prs and VT appeared highest and the Ti longest at the optimal initial PS How setting. min. This compares to reported fixed initial PS flows in other ventilators that range from 58 to 100 Umin, depending upon manufacturer. 2 Examples of airway pressure tracings during maximal, optimal (ie, the initial flow setting with the highest PV value) and minimal flow setting are given in Figure 2. Note the marked variations in delivered pressure and volume that could be observed over this range of initial PS flow settings. The PV product and the attained PPS as a function of initial PS flow are plotted for all 17 patients in Figure 3. Note from these plots that the ability of the ventilator to reach PSSEL is reduced when initial PS flow is not optimal, especially when initial PS flow is less than optimal (p<o.o when comparing PPS at optimal to that at maximal or minimal settings). The ventilatory patterns (VT, f, and T)as a function of initial PS flow are plotted for all 17 patients in Figure 4. Note from these plots that VT is generally largest, f is generally slowest, and Ti is generally longest at the optimal initial PS flow setting (p<o.o when comparing the optimal VT to either maximal or minimal VT, p<o.02 when comparing the optimal f to eithermaximal or minimalf, p<o.oowhen comparing optimal T, to either maximal or minimal T). This is consistent with the concept that, for a given level of PS (ie, PSSEL), the patient's ventilatory control system at the optimal initial PS flow setting will hold the inspiratory effort longer and allow the ventilator to deliver a largervtat its set pressure than at suboptimal flow settings. Linear regression analysis showed significant correlations between the optimal initial PS flow setting and indices of ventilatory loads and ventilatory drive (Table 2). These correlations are consistent with the concept that the greater the load and the more intense the drive, the faster the flow that will be needed to synchronize with the patients ventilatory demands. Varying Breath Termination Criteria The response to changing breath termination criteria in the 16 patients was minimal. Although T significantly decreased as breath termination criterion was changed from 25 to 50 percent peak Bow; there was no significant effect on the VT,f, or delivered Prs (Table 3). No evidence of increased expiratory work (ie, terminal elevations in inspiratory airway pressure) was observed at the longest breath termination setting (25 percent of peak flow). DSCUSSON Several investigators have demonstrated that the spontaneous inspiratory effort which triggers a mechanically assisted breath does not cease with the delivery of that breath,"?" Thus, the potential for significant imposed ventilatory muscle loads exists if the mechanically assisted flows and volumes do not match the patient's desired flow pattern and tidal volume. Subjective dyspnea, as well as increased muscle energy demand, would appear to be a conse- 136 nilial FlowRate,B8athTermination Criteria,and PressureSUpport Ventilalion (Macntyre. Ho)

4 ~30...J <:& :J: 20 E ~ > Q ~1000. ~ > L. J L l - - l l. - J - : t. = : t : : = : : Siower- Optimal -Faster nitial PS Flow Setting FGURE 3. The ventilator pressure volume product (PV-fop panel) and airway pressure (Pes as a fraction of PSSEL-bottom panel) as a function of delivered initial PS flow. Delivered initial PS flow is expressed in relation to the optimal initial PS flow (ie, that flow associated with a maximal PV product). Values are thus plotted at the optimal initial PS flow and then at the other six flow settings that were either above or below this optimal setting. Note that the PV product (by definition) was highest at the optimal initial PS flow setting and that it could drop markedly at flows both above and below this. Note also that at the optimal initial PS flow setting, PPS was closest to PSSEL and that, especially at flows below the optimal, PPS could be markedly lower than PSSEL. quence of such imposed loads,"-" This dyssynchrony has been shown to be an important source of imposed load when using volumeassisted breaths because of the fixed flow and volume characteristics of such breaths.>" Moreover, if these volume-assisted breaths are given only intermittently, there appears to be a consequent increase in patient ventilatory drive thatcan further increase this imposed load." n contrast, the flow, timing and volume pattern of PS is designed to adjust to patient demand with every breath.y Because of this, the dyssynchrony Table 2-unear Regression &sults Comparing Optimal nitial PS Flow Seuings to ndices of Ventilation Wad and Ventilatory Drive Slope ntercept r p nitial PS Flow vs: CRS <0.01 MV Not significant PTRG <0.01 u $! i= 1.0 Breath Termination 25% Peak Flow 50% Peak Flow Criterion (Mean ± SD) (Mean ± SD) p Pps, cm H 2O 23± ±6.0 NS PTRG, cm H 2O -2.5±1.l5-2.8±0.69 NS n, s 1.l± ±0.2 <0.05 VT,ml 800± ±166 NS f, breaths/min 18.8 ± ±7.8 NS CHEST 99 1 JANUARY,

5 flow pattern that was most synchronous with a spontaneous flow demand would result in the largest VTat a pressure closest to the selected PS (ie, highest ventilator "work" or "output"). Given this definition of optimal initial PS flow, our results would suggest the following patient ventilator interactions during high level PS (ie, levels of PS sufficient for a VT of 9 to 12 mjjkg). First, initial PS flows below the optimal result in progressive reductions in VT, increases in f, decreases in nand, eventually, nonattainment of the selected PS level. This response is consistent with the normal ventilatory pattern changes that occur with increasing ventilatory loads as described by others Second, initial PS flows above the optimal also result in progressive reductions in VT, increases in f and decreases in Ti. Two explanations for this exist: (a) since PS flow termination criteria are usually functions of initial PS flow (ie, 25 percent), inappropriately high initial PS flow may cause premature flow termination for a given patient effort; (b) alternatively, overly rapid airway pressurization may cause premature termination of the actual patient's inspiratory effort. Our data do not allow us to distinguish between these two mechanisms in a given patient. Third, the optimal initial PS flow appears to correlate with load and drive (ie, the higher the load and the stronger the drive, the faster is the optimal initial PS flow). This is consistent with observations that have shown an increasedspontaneousflow demand as ventilatory loads increase.p'" Varying breath termination criteria appear to have a lesser effect. Most ventilators terminate PS when the inspiratory flow decreases to 25 percent of the peak inspiratory flo\. 2 Although exhalation is generally passive, active expiratory work of breathing can occur. Usually, this is a result of external resistance of the ventilator circuitry (eg, flow resistance of expiratory valve apparatus16) or the degree ofasynchrony between patients respiratory pattern and the ventilator cycle. 17 We speculated that fixed PS termination criteria may also compromise patient-ventilator synchrony. Our results, however, do not show any significant difference in tidal volume or respiratory rate when PS termination criterion was changed from 25 to 50 percent of peak inspiratory flow t thus appears that this range of adjustment is unimportant in optimizing patientventilator synchrony. Preliminary reports on reducing PS termination criterion to 5 percent of peak flo\v, however, do show effects on the ventilatory pattern,17 and thus, further study on this subject is warranted. n summary, these results suggest that, in addition to the level of applied PS, the initial flow to reach the PS level is also important in providing optimal PS. Specifically, previous studies have clearly shown that the level ofps correlates with the degree ofventilatory muscle unloading and consequent changes in the ventilatory pattern. 1,3,4,18 Now it appears that for this unloading effect to be optimally synchronized with patient demands, an adjustable initial flow capability is needed in some patients. On the other hand, altering breath termination criterion from 25 to 50 percent of peak initial flow appears to have minimal effect. ACKNOWLEDGMENT: The authors are indebted to Mrs. Janet Johns for her secretarial expertise. REFERENCES 1 Macintyre NR. Respiratory function during pressure support ventilation. Chest 1986;89: Kacmarek RM. The role of pressure support ventilation in reducing work of breathing. Respir Care 1988;33: Macintyre NR, Leatherman NE. Ventilatory muscle loads and the frequency-tidal volume pattern during inspiratory pressure assisted (pressure supported) ventilation. Am Rev Respir Dis 1990; 141: Brochard L, Pluskwa F, LeMaire F. mproved efficiency of spontaneous breathing with inspiratory pressure support. Am Rev Respir Dis 1987; 136: Macntyre N, Nishimura M, Usada Y, et al. The Nagoya conference on system design and patient-ventilator interactions during pressure support ventilation. Chest 1990;97: Toldoka H, Saito S, Kosalca F. CompaJison of pressure support ventilation and assist control ventilation in patients with acute respiratory failure. ntensive Care Med 1989; 15: Marini JJ, Capps JS, Culver BH. The inspiratory work of breathing during assisted mechanical ventilation. Chest 1985; 87: Marini JJ, Smith TC, Lamb VJ. External work output and force generation during synchronized intermittent mandatory ventilation. Am Rev Resp Dis 1988; 138: Sassoon CSH, Mabutte CK, Te TT, Simmons DH, Light Rvv. Work of breathing and airway occlusion pressure during assist mode mechanical ventilation. Chest 1988; 93: Ward ME, Corbell C, Gibbons \Y, Newman S, Macklem PT. Optimization of respiratory muscle relaxation during mechanical ventilation. Anesthesiology 1988;69: DerenneJP, Macklem PT, BoussosC. The respiratory muscles: mechanics, control and pathophysiology (part 2). Am Rev Respir Dir 1978; 118: Tobin MJ, Perez \Y, Guenther SM, Semmes BJ, et al. The pattern of breathing during successful and unsuccessful trialsof weaning from mechanical ventilation. Am Rev Respir Dis 1986; 134: Milic-Emili. Recent advances in clinical assessment of control of breathing. Lung 1982; 160: Agostoni E, D'Angelo E, Piolini M. Breathing pattern in men during inspiratory elastic loads. Respir Physioll978; 34: Aubier M, Murciano D,Foumier M, Milie-Emili J, Pariente R, DerenneJP.Central respiratory drive in acute respiratory failure of patients with chronic obstructive pulmonary disease. Ain Rev Respir Dis 1980; 122: Banner MJ, Downs JB. Effects of expiratory flow resistance on inspiratory work of breathing. Chest 1988;93: Smith TC, Marini JJ, Lamb VJ. Active expiratory work of breathing during mechanical ventilation [abstract]. Am Rev Respir Dis 1986; 133: Braschi A, Sala-Gallini G, RodiG, lotti, Chiaranda M, Villa S. Relationships between sensitivity of the expiratory trigger and breathing pattern during pressure support ventilation (abstract). Am RevRespjr Dis 1989; 139: Brochard L, Harf A, Lorino H, LeMaire F. Pressure support preventsdiaphragmatic fatigue during weaning from mechanical ventilation. Am Rev Respir Dis 1989; 139: nltiaj FlowRate. 88ahTermination Crieria, and Pnlssu8 Support Venlialon (Mac/nlyre. Ho)