Pressure-Time Product during Continuous Positive Airway Pressure, Pressure Support Ventilation, and T-Piece during Weaning from Mechanical Ventilation 1- 3

CATHERINE S. H. SASSOON, RICHARD W. LIGHT, ROMI LODIA, GARY C. SIECK, and C. KEES MAHUTTE

Introduction

T he goal of weaning from mechanical ventilation is to have the patient assume spontaneous breathing with maintenance of adequate gas exchange. T-piece, continuous positive airway pressure (CPAP), and pressure support (PS) are ventilatory modes used to achieve this goal (1-5). With these weaning modalities, the load that the inspiratory muscles must overcome varies. The levelof inspiratory muscle work (WI)during weaning is obviously an important determinant of weaning (6, 7). If the inspiratory muscles have to work too hard, fatigue will ensue and weaning will fail. On the other hand, if the respiratory muscles are rested completely for too long, they lose strength (8). The optimal balance between WI and rest during weaning has yet to be determined. Likewise, there is relatively little information comparing the WI during various weaning modalities. The reported effects of CPAP on WI have been conflicting (9-16). The effect of CPAP on WI will depend on the subject's respiratory system mechanics and dynamic hyperinflation with its associated intrinsic positive end-expiratory pressure (PEEPi). The reported discrepancies in the literature may be due to these factors. PS would be expected to decrease WI. The extent of the reduction would be proportional to the level of preset pressure at relatively low levels of PS. One previous study (17) showed that PS of ~ 10 em H 2 0 decreased WI, but in one patient the WI was least with a PS level of zero em H 2 0 . Little information is available on the effects of 5 em H 2 0 PS on WI. The objective of the present study was to compare the pressure-time product (PTP), PEEPi, and lung mechanics during weaning with low levelsof CPAP, PS, and T-piece. In this study we used the

SUMMARY The objective of this study was to compare the effects of continuous positive airway pressure (CPAP), pressure support ventilation (PS), and T.plece on the pressure-time product (PTP) during _anlng from mechanical ventilation. The PTP Is an estimate of the metabolic work or oxygen consumption of the respiratory muscles. We studied 10Intubated patients recovering from acute respiratory failure of verlous etiologies. A modified continuous flow (flow-by) CPAP of 0 and S cm H20 (CPAP-oand CPAP-S, respectively), PS of S cm H20 (PS-S),and T-plece _re applied In random order for 30 min each. In the last S min of the 30-mln periods, _ measured the esophageal pressure and tranadlaphragmatlc pressure·tlme prodUCts- PTP(es)and PTP(dl), cm H20'slmln, respectlvelymultiplied by respiratory frequency. Breathing pattern, total lung resistance (RL), quasi-static lung compliance (CL), Intrinsic positive end-explratory pressure (PEEPI), end-explratory transpulmonary pressure (Plpexp), arterial blood gases, blood pressure, and heart rate _re also measured. In comparison to T-plece, CPAP-S decreased PTP(es) 40% (p < 0.01)and PTP(dl) 43% (p < 0.02), whereas PS-Sdecressed PTP(es) 34% (p < 0.01)and PTP(dl) 38% (p < O.OS). The decrease In PTP(es) with CPAP.Swas associated with a significant reduction In RL, and to a less extent In PEEPIrelative to airway pressure. The contribution of the decrease In PEEP' to the reduction In PTP(es) amounted to 36%. With PS-S, respiratory system mechanics and PEEPI ware not significantly different eempared with T-plece. With CPAP-O, PTP tended to be lowsr than with T-plece. This was likely due to the Inherent small end-Inspiratory positive pressure related to the CPAP system. Breathing pattern, arterial blood gss measurements, and hemodynamics were similar In all the ventilatory modes. In conClusion, during _anlng from mechanical ventilation of patients recovering from acute respiratory failure, low levels of CPAP and PS significantly decrease the pressure-time product by 40 and 34%, respectively, compared with T.plece. AM REV RESPIR DIS 1991; 143:469-475

PTP as an estimate of the metabolic work or the oxygen consumption of the respiratory muscles (V02resp) (18-20). Methods Subjects We studied 10 male patients during weaning from mechanical ventilation. The clinical characteristics of these patients are listed in table 1. For inclusion in this study, the patients had to be clinically and hemodynamically stable and judged ready to wean by their primary physician. They had to have a maximal inspiratory pressure(MIP) of at least - 20 em H 20 or a FVC of at least 10 ml/kg body weight. One patient (Patient 1) was studied twice. During the first study, the patient developed acute respiratory acidosis during CPAP of zero em H 2 0 used as the initial weaning method. The study had to be terminated, and the data were not included in the analysis.The patient consented to be restudied 4 days later after stabilization of the acute respiratory failure. Each patient or his next

of kin signed a written informed consent form approved by the Institution's Human Experimentation Committee. Protocol The study was conducted in a semirecumbent position (f\J 30 degrees from the horizontal position). Wemeasured MIP with a pressure transducer (Valydine MP-45 ± 100em H 2 0 , (ValydineCorp., Northridge, CAl by having

(Received in original form October 19, 1989 and in revised form October 10, 1990) I From the Departments of Medicine, Veterans Administration Medical Center, Long Beach, the University of California, Irvine, California, and the Departments of Anesthesiology, Physiology, and Biophysics,Mayo Clinic, Rochester,Minnesota. 2 Supported by the Department of Veterans Affairs Medical Research Service. 3 Correspondence and requests for reprints should be addressed to Catherine S. H. Sassoon, M.D., Pulmonary Section (llIP), Veterans Administration Medical Center, Long Beach, CA 90822.

469

470

SASSOON, LIGHT, LOOIA, SIECK, AND MAHUTTE

TABLE 1 SUBJECTS' CHARACTERISTICS

Patient No .

1 2 3 4 5 6 7 8 9 10 Mean ± SO

Age

(yr)

70 69 79 40

Weight (kg)

rr

44.6 84 .5 125.0 65.9 62.0

63 47 66 73 67 65 12

126.0 56.2 74.3 61.0 74.7 77.4 26.2

MIP (em H2O)

FVC (L)

-33

0.50

-36 -110

NA NA NA

-40

1.10

-59 -68 -20 -42 -58 -51 24

1.39

-40

NA

0.78 1.00 0.80 0.92" 0.31

ET Tube (size)

8.0 8.5 8.5 8.5 8.0 7.5 8.5 8.0 7.5 8.5

Durat ion of AV (days)

13 9 9 4 34

1 7 2 3 3 9 10

Etiology of Respiratory Failure COPO COPO, alcoholic cardiomyopathy Pneumonia, acute renal failure, obesity Pulmonary edema, chronic renal failure Pulmonary edema , myocardial infarction, pneumonia, COPO COPO, obesity Asp irat ion pneumonia COPO COPO CO PO

Definition of abbreviations: MIP = maximal inspiratory pressure; FVC = forced vital capacity; ET = endotracheal tube; AV assisted ventilation; COPD = chronic obstructive pulmonary disease; NA = not available. • n - 6.

the patient inspire maximally against an inspiratory occlusion at end-expiratory lung volume held for 1 s. In the four patients (Patients 2, 3, 4 and 7) who could not follow commands, the MIP was obtained once by occluding the inspiratory line until a maximal airway pressure plateau was achieved(21). The FVC was measured with a Boehringer spirometer (Boehringer Laboratories, Wynnewood, PA). Three maneuvers of 1-sMIP and FVC were performed, and the largest values are reported. For 30 min each the patients were then placed on four different modes of ventilation in random order: (1) CPAP of 0 ern H 20 (CPAP-O), (2) CPAP of 5 em H 20 (CPAP-5), (3) pressure support ventilation of 5 em H 20 (PS-5), and (4) T-piece. The CPAP and PS ventilations were delivered via the Puritan-Bennett nooa ventilator (Puritan Bennett Corp., Carlsbad, CA). To minimize the contribution of the CPAP system to inspiratory muscle work, we used a modified continuous flow system (flow-by) as previously described (22). The base flow was set at 10 L/min with a flow sensitivity of 2 to 3 L/min. With PS, the ventilator sensitivity was set at -I em H.O. During the study, the inspired oxygen fraction (Flo 2 ) was maintained the same as during full ventilatory support. However, in four patients (Patients 2, 3, 6, and 7) who had an Flo. between 25 and 35010 during assisted mechanical ventilation, Flo. had to be increased to 40% when breathing through a T-piece because the minimum delivered Flo. on our Tpiece system was 40%. At the completion of the study, all but one patient was extubated and received supplemental oxygen (02 ) via a Venturi Mask with an Flo. similar to that during assisted ventilation. The exception was Patient 5 who had a tracheostomy and wastherefore maintained on a T-piece. All patients sustained spontaneous breathing within 24 h of extubation. In the last 5 min of the 30-min period, the

=

following variables weremeasured and/or calculated: the esophageal and transdiaphragmatic pressure-time products - PTP(es) and PTP(di), respectively-tidal volume (VT), respiratory frequency (f), total minute ventilation (VE) as the product of VTand f, inspiratory and expiratory time (Tt and Th, respectively), total breath cycleduration (not), the ratio of Tr/Ttot and Th/Ttot, mean inspiratory and expiratory flow rate (VTIn and VT/Th, respectively), changes in esophageal pressure (aPes) and transdiaphragmatic pressure (apdi), end-expiratory transpulmonary pressure (Ptpexp), intrinsic positive endexpiratory pressure (PEEPi), total lung resistance (RL), quasi-static lung compliance (CL), heart rate (Fe), blood pressure (BP) and arterial blood gases. To ascertain whether there were concomitant changes in the electrical activity of the diaphragm (EMGd) as PTP was altered with all the ventilatory modes, in five patients, we also measured the EMGd using bipolar esophageal electrodes. The breathing circuit consisted of the endotracheal tube, a heated pneumotachograph (Fleisch No.2; Fleisch, Lausanne, Switzerland), and either the ventilator circuit or the T-piece tubings. Esophageal pressure (Pes) (which reflects pleural pressure) was measured with an esophageal balloon catheter connected to a differential pressure transducer (Valydine MP45 ± 50 em H.O) using the standard technique (23). Transdiaphragmatic pressure (Pdi) was measured using electronic subtraction of Pes from the gastric pressure (Pga), The latter was measured with a balloon catheter filled with 2 ml of air placed in the stomach and a differential pressure transducer (Valydine MP45 ± 50 em H.O) . Tho patients refused insertion of the gastric balloon. The PTP(es) was measured as the area subtended by Pes and the chest wall static recoil pressure (Pstw)-time curve over Tt, taking into account the PEEPi (figure I). The Pstwtime curve was extrapolated from the Pstwvolume curve of normal subjects, making the

assumption that these relationships were linear within the tidal volume range. The static compliance of the chest wall (Cstw) of normal subjects in the supine position amounted to 5% of predicted vital capacity per ern H.O (24). Thus, the slope of the Pstw-time relationship = (aVT/Cstw)1 a 11. The PTP(di) was measured as the area subtended by Pdi above its expiratory baseline over Tt (figure I). The areas of interest for PTP(es) and PTP(di) were measured with a digitizer (Sigma Scan vers 3.01; Jandel Scientific, Corte Madera, CA). For PTP(es) and PTP(di), the average values of three breaths or more were used for analysis. Both PTP(es) and PTP(di) (n = 8) were multiplied by respiratory frequency and expressed as em H 20·s/min. Airway pressure (Paw) was measured through a port proximal to the endotracheal tube with a differential pressure transducer (Valydine MP45 ± 100 em H.O). Flow (V) was measured with a heated pneumotacho-

P slw - I lm e c u r v e

Pes

Pdl

TI ME

Fig. 1. Graphic measurements of the PTP(es) and PTP(di) per breath as estimates of i nspiratory metabol· ic work. PTP(es) is the horizontal and the cross-hatched areas subtended by the esophageal pressure (Pes), intrinsic end-exp iratory pressure (PEEP i), and chest wall static recoil pressure (Pstw)-timecurve over 11. The crosshatched area is the additional work induced by PEEP i. PTP(dl) Is the horizontal-hatched area subtended by the transdiaphragmatlc pressure (Pdi)above end-expiratory baseline over 11 (iI = flow; insp '" inspiration; exp = expiration; VT = tidal volume). See text for further explanation .

PRESSURE-TIME PRODUCT DURING CPAP, PRESSURE SUPPORT, AND T-PIECE

471

measured (figure 3), the correlation between PTP(di) and PTP(es) was significant for all ventilatory modes (r = 0.90, -;; 300 P < 0.001; n = 32) . Although PTP(es) e and PTP(di) were also lower with CPAP-5 Fig. 2. PTP(es) (ern H20·slmin) during o and PS-5 compared with CPAP-O, these CPAP-O and CPAP·5. PS-5. and T-piece. ~'" 200 Asterisks indicate p < 0.01 CPAP-5 and differences did not achieve statistical o PS-5versus T-piece; n = 10;closed ctrsignificance. cles = mean. It can be seen in table 2 that the breath" 100 Q. ing pattern did not differ significantly Q. among the ventilatory modes. During CPAP-5 and PS-5, ~Pes was significantly O -l-----, ---~----,----~lower compared with that during T-piece, CPAP -O CPAP -S ps-S T- piece and it tended to be lower than that during CPAP-o. Similarly, ~Pdi was significantly lower during CPAP-5 than during T-piece, graph and a differential pressure transducer artery and a disposable pressure transducer and during PS-5 than during CPAP-O. PEEPi in individual subjects varied (Valydine MP45 ± 2 em H 2 0 ), VT was ob- (MK5-04DTNVF Sorenson Transpac U; Abtained by electronically integrating the flow bott Laboratories, North Chicago, IL). Mean from breath to breath. However, in 28 signal (8815A; Hewlett-Packard, Waltham, BP was used for analysis. Arterial blood was (70%) of 40 PEEPi measurements (10 paMA). TI, Ts, not, and f werecalculated from withdrawn from the arterial catheter and ana- tients, each with four different weaning the flow signal. Ptpexp wascalculated by sub- lyzed with a blood gas analyzer (Model 1306; modalities), the difference between the tracting Pes from Paw at end-expiration. ~Pes Instrumentation Laboratories, Lexington, MA). highest and lowest values of PEEPi in and ~Pdi were calculated as the change in For EMGd measurement, stainless steel individual subject was ~ I cm H 0 . In 2 pressure from end-expiration to end-inspira- bipolar electrodes were affixed to the gastric the remainder (12 of 40), these differences tion of the Pes and Pdi, respectively. End- balloon catheter. For each mode of ventilaexpiration and end-inspiration were deter- tion, the catheter was positioned to provide ranged between 1.2 and 2.4 em H 2 0 . Almined from the flow tracing at po ints of zero the highest signal to noise ratio. The EMGd though in our ready-to-wean patients, flow. All signals were recorded on an eight- signal was amplified (BMA 831; CWE, Ard- mean PEEPi during all ventilatory modes channel recorder (No. 7758B; Hewlett- more, PA) and filtered between 10and 1,000 was small, PEEPi decreased from a mean Packard). The above variables were calculat- Hz. The duration ofEMGd activity wasquan- of 3.5 em H 2 0 during T-piece (range, 1.3 ed as the average values for 10 breaths. titated from the onset to the cessation of the to 8.2 em H 2 0 ) to 2.3 em H 2 0 during RL was calculated using the isovolume electrical activity. CPAP-5 (range, l.l to 3.5 em H 2 0 ; p < method (25). Changes in Pes subtracted from 0.01). Two patients (Patients 9 and 10) Statistical Analysis Paw were divided by changes in V at midinhad PEEPi greater than 5 em H 2 0 while spiratory and midexpiratory lung volumes.CL Means and standard errors were calculated breathing on T-piece. In these patients, was calculated as the ratio of the change in for each variable, unless otherwise indicatlung volume to the change in Pes subtracted ed. Analysis of variance was used for com- PEEPi decreased from 5.3 to 3.5 ern from Paw at points of zero flow (26). In this paring means. If the F value was significant, H 2 0 , and from 8.2 to 3.1 em H 2 0 durregard, PEEPi did not enter in the calcula- the least significant difference between means ing CPAP-5, respectively. It should be tion of CL. PEEPi was measured from the was calculated. Where applicable, regression pointed out that the decrease in PEEPi Pes tracing as the deflection from the plateau analysis was used; probability values lessthan during CPAP was relative to airway prespressure during exhalation to the point when 0.05 were considered significant. sure rather than to atmospheric presinspiratory flow began (27) with the paper sure. As expected, Ptpexp increased with speed set at 100mm/s, For measurements of Results CPAP-5 (p < 0.01). During CPAP-5, RL RL, CL, and PEEPi, the average values of Both CPAP-5 and PS-5 decreased PTP(es) .: decreased significantly, whereas CL inthree breaths were used for data analysis. Heart rate (Fe)wasmonitored with an elec- and PTP(di) significantly, compared with ' creased compared with T-piece. However, trocardiogram. BP was directly measured T-piece (figures 2 and 3, respectively). In the latter did not achieve statistical significance. During CPAP-5, CL decreased using an arterial catheter placed in the radial the eight patients in whom PTP(di) was in four of the patients (Patients 4, 5, 8, and 9) and increased in the other six patients compared with T-piece. SOO Arterial blood gas determinations, BP, c and Fc were not significantly different ,E 400 (table 3) during all ventilatory modes. o e The analog tracings of pressures V, VT, Fig. 3. PTP(di) (cm H20 ·slmin) during 6 300 and EMGd of a representative subject CPAP-