Inflation and deflation timing of the AutoCAT 2 WAVE intra-aortic balloon pump using the autopilot mode in a clinical setting

Transcription

1 450060PRF / Bakker EWM et al.perfusion 2012 Inflation and deflation timing of the AutoCAT 2 WAVE intra-aortic balloon pump using the autopilot mode in a clinical setting Perfusion 0(0) 1 6 The Author(s) 2012 Reprints and permission: sagepub. co.uk/journalspermissions.nav DOI: / prf.sagepub.com EWM Bakker, 1 K Visser, 1 A van der Wal, 1 MA Kuiper, 2 M Koopmans 2 and R Breedveld 3 Abstract The primary goal of this observational clinical study was to register the occurrence of incorrect inflation and deflation timing of an intra-aortic balloon pump in autopilot mode. The secondary goal was to identify possible causes of incorrect timing. During IABP assistance of 60 patients, every four hours a strip was printed with the IABP frequency set to 1:2. Strips were examined for timing discrepancies beyond 40 ms from the dicrotic notch (inflation) and the end of the diastolic phase (deflation). In this way, 320 printed strips were examined. A total of 52 strips (16%) showed incorrect timing. On 24 of these strips, the incorrect timing was called incidental, as it showed on only one or a few beats. The other 28 cases of erroneous timing were called consistent, as more than 50% of the beats on the strip showed incorrect timing. We observed arrhythmia in 69% of all cases of incorrect timing. When timing was correct, arrhythmia was found on 13 (5%) of 268 strips. A poor quality electrocardiograph (ECG) signal showed on 37% of all strips with incorrect timing and 11% of all strips with proper timing. We conclude that inflation and deflation timing of the IABP is not always correct when using the autopilot mode. The quality of the ECG input signal and the occurrence of arrhythmia appear to be related to erroneous timing. Switching from autopilot mode to operator mode may not always prevent incorrect timing. Keywords intra-aortic balloon pump; IABP; timing; deflation; inflation; AutoCAT 2 WAVE; autopilot Introduction Since the introduction of the intra-aortic balloon pump (IABP) in cardiac care, it has been applied as an effective method to improve left ventricular performance. 1,2 Currently, it is the most frequently employed type of cardiac assist device, used to counter a period of ischemic left ventricular failure. Effective use of the IABP depends on accurate inflation and deflation timing. 3 Late inflation and early deflation lead to suboptimal assistance, where early inflation and late deflation may augment the afterload, thereby, causing an increase in cardiac oxygen consumption and impaired left ventricular performance. 4 Some new methods have been introduced in the last decade for determining the points of inflation and deflation. The fibre-optic sensor provides a reliable pressure signal. 5 From this signal the aortic flow is calculated on a beat-to-beat basis, providing the aortic valve closure point for that beat. With the use of this sensor, new algorithms have been applied to focus on accurate detection of the dicrotic notch 6,7 and better handling of arrhythmias. 8 Furthermore, trigger timing and trigger choice may change without human intervention as the autopilot accurately monitors the effects of the IABP on patient parameters. The autopilot ensures a timely change to the arterial pressure trigger 1 Departement of Perfusion, Medisch Centrum Leeuwarden, 2 Departement of Intensive Care, Medisch Centrum Leeuwarden, 3 Departement of Cardiac Care, Medisch Centrum Leeuwarden, Corresponding author: Edwin Bakker Dept. of Extracorporeal Circulation Hart Centrum, Medisch Centrum Leeuwarden H. Dunantweg 2, 8901 BR Leeuwarden edwin.bakker@znb.nl

2 2 Perfusion 0(0) when the ECG signal is lost, and back when it is regained. It may also change the algorithm used, for example, when arrhythmias are detected. With these changes the need for human intervention decreases. But this advantage may have its drawbacks. When the IABP function is entrusted to the autopilot, timing errors due to haemodynamic changes, poor quality trigger signals or incorrect position of the catheter could be missed by the nursing staff. The new algorithms have proven to perform well, 6,8,9 but the reliability of the autopilot, in all of its facets, has not been extensively examined in a clinical setting. Osentowski 10 describes the performance of the autopilot when using the Datascope CS100 IABP (Maquet, Hirrlingen, Germany). Incorrect timing could be demonstrated in 16.7% and 55.6% of inflations and deflations, respectively. We hypothesized that inflation and deflation timing of the IABP may not always be correct using the autopilot mode of the IABP in a clinical setting. Therefore, we investigated the timing of the Arrow AutoCAT 2 WAVE (Windkessel Aortic Valve Equation) IABP with 60 patients requiring IABP assistance. The input signals of ECG and blood pressure may present with arrhythmias or artefacts, possibly caused by other clinical equipment or movement of the catheter within the aorta. 14 Therefore, the second aim of this study was the identification of the causes of erroneous timing, wherever possible. Methods During a period of 29 months ( ), the inflation and deflation timing of the AutoCAT 2 WAVE IABP (Teleflex Incorporated, Hilversum, ) was tested while using the autopilot mode. The IABP timing was observed during IABP assistance of 60 patients on the cardiac care unit or the intensive care unit of the Medical Centre Leeuwarden. In 11 of these cases, observation started with a patient in the operating theatre. The IABP was supplied with software update 2.23 (SPSS Inc., Chicago, IL, USA), ensuring optimal handling of arrhythmias, breathing artefacts and trigger changes. This update also changed the timing of inflation from 20% to 15% of maximal (calculated) flow, which would decrease the occurrence of (borderline) early inflation. A 40-cc, 7.5 Fr. UltraFlex TM IAB-Catheter (Teleflex Incorporated) with fibre-optic sensor was used in 39 patients. This type of catheter was recalled in The reason for the recall was that some catheters were not recognized by the IABP console, resulting in an inflation volume of 5 cc instead of 40-cc. As a result of this recall, a 40-cc, 8 Fr. Fidelity TM IAB-catheter (Maquet, Getinge, Sweden), without a fibre-optic sensor, was used in 19 patients. A 40-cc, 8 Fr. NarrowFlex IAB- catheter (Teleflex Incorporated), without a fibre-optic sensor, was used in two patients. Catheters were inserted through the femoral artery. The IABP was used in the autopilot mode. Every four hours of assistance, the IABP frequency was temporarily set to 1:2 and a strip was printed from the IABP console, presenting the ECG and the aortic blood pressure waveform of at least five heart beats. ECG leads from the patient led to the IABP console directly. Where possible, blood pressure was measured by the fibre-optic sensor. In the operating theatre, the ECG signal was slaved from the monitor. The central lumen of the catheter was used as the blood pressure signal source when a fibre-optic sensor was not present. Factors that might affect the inflation or deflation timing were written down by the nursing staff. The strips were independently examined by two perfusionists and an intensivist for incorrect timing of the inflation or deflation, according to the following guidelines: 1) inflation should be less than 40 ms (1 mm) from the dicrotic notch, 2) deflation should be less than 40 ms from the start of the next systole and should visibly reduce end diastolic pressure. These guidelines were chosen after consulting the company (Teleflex) about the deviation justifying a switch from autopilot mode to operator mode. The number of early or late inflations and early or late deflations were noted. Incorrect timing was called consistent if it showed on >50% of all beats of a printed strip. Only one or a few incorrectly timed beats (with a maximum of 50% of all beats) on a print was called incidental. Results We included 320 from 339 printed strips. Eighteen strips were excluded, as the IABP was changed from autopilot mode to operator mode. The reasons for changing the mode were late deflations (four times), early deflations (three times) and early inflations (once). In 10 cases, the reason for conversion to operator mode was not identified. One strip showing a very poor quality blood pressure curve was excluded. Consistent incorrect timing was seen on 28 strips (9%) from 15 patients. As seen in Table 1, incorrect deflation appeared on 21 of these strips where, in seven cases, inflation was not correct. Examining possible causes for erroneous timing, we found incorrect timing to occur when arrhythmia and/ or distorted ECG signals were observed (Table 2). Incorrect positioning of the catheter on X-thorax was seen in combination with distorted ECG signals in three cases. On three strips, we found late inflations that may relate to haemodynamics (short systole). The quality of the pressure curve was poor on one strip

3 Bakker EWM et al. 3 Table 1. Incidence of consistent incorrect timing of the IABP Type of incorrect timing Early inflation 4 (1%) Late inflation 3 (1%) Incorrect inflation 7 (2%) Early deflation 11 (3%) Late deflation 20 (6%) Incorrect deflation 21 (7%) * Total 28 (9%) number of strips (n=320) *The total number of strips with incorrect deflation timing was lower than the sum of early and late deflations because both types of incorrect deflation were seen on 10 strips. IABP: intra-aortic balloon pump. with incorrect timing. Incorrect timing appeared not to be related to the type of catheter used, although damped pressure curves were not seen when using a FOS signal. On 24 strips (7%), we found incidental incorrect timing, 21 concerning deflation, three with erroneous inflation timing. During all cases of incidental incorrect timing, arrhythmia or incidental irregular beats or artefacts of the ECG signal were observed, as seen in Table 3. In one case of early inflation, we found two augmentations during systole, as shown in Figure 1. Arrhythmia was found in 36 cases of the 52 strips that showed consistent or incidental incorrect timing (69%). In 13 cases (5%) of all 268 strips with correct timing, an irregular rhythm was found (Table 4). In 19 of the 52 strips with incorrect timing (37%), ECG quality was poor. Of all strips with correct timing, 30 out of 268 strips (11%) presented with a poor quality ECG, as shown in Table 4. The quality of the ECG was determined in two ways. The AutoCAT 2 IABP has a quantitative expression of the ECG quality printed on the strip ( ECG score in Table 2 and Table 3). In 16 cases, the quality score of the chosen lead was four or higher, which means that the ECG quality is poor. In 33 cases, the ECG was significantly disturbed, in spite of a good ECG quality score (an example is seen in Figure 1). Table 2. Consistent incorrect timing and observations Strip nr Inflation Deflation Observations Cath Trigger (algorithm) ECG score 34 late Short systole NF ECG, skin (pattern) 0 35 late Short systole NF ECG, skin (pattern) 0 36 late Short systole NF ECG, skin (pattern) 0 40 late - NF ECG, skin (pattern) 0 48 late APqual NF ECG, skin (pattern) 0 63 late+early ECGqual FOS ECG, skin (pattern) 0 71 early - FOS ECG, skin (pattern) 0 73 early ECGqual FOS ECG, skin (pattern) late+early Arrhythmia FOS ECG, skin (peak) late Arrhythmia FOS ECG, skin (pattern) late+early Arrhythmia/ECGqual NF ECG, skin (pattern) late+early Arrhythmia NF ECG, skin (peak) late+early Arrhythmia/ECGqual NF ECG, skin (peak) early Arrhythmia NF AP late Arrhythmia NF ECG, skin (pattern) late+early Arrhythmia/ECGqual NF ECG, skin (peak) late POS/ECGqual NF ECG, MON (peak) late POS/ECGqual NF ECG, MON (pattern) late POS/ECGqual NF ECG, skin (pattern) late ECGqual NF ECG, skin (pattern) early FOS ECG, MON (pattern) early FOS ECG, MON (pattern) late Arrhythmia FOS ECG, skin (peak) late+early Arrhythmia FOS ECG, skin (peak) late Arrhythmia/ECGqual FOS ECG, skin (peak) late+early Arrhythmia/ECGqual FOS ECG, skin (pattern) late+early Arrhythmia FOS ECG, skin (peak) late+early Arrhythmia FOS ECG, skin (peak) 0 APqual: poor quality arterial pressure signal, ECGqual: poor quality ECG signal, POS: position of the IAB catheter too low, Cath: catheter type, FOS: catheter with fibre-optic sensor, NF: catheter without fibre-optic sensor, ECG skin: ECG signal directly from patient skin, ECG MON: ECG signal slaved from monitor, AP: arterial pressure signal.

4 4 Perfusion 0(0) Table 3. Incidental incorrect timing and observations Strip nr Inflation Deflation Observation Cath Trigger (algorithm) ECG score 143 late Arrhythmia FOS ECG, skin (pattern) late Arrhythmia FOS ECG, skin (pattern) late+early Arrhythmia FOS ECG, skin (pattern) late Arrhythmia/ECGqual NF AP early Arrhythmia/ECGqual NF ECG, skin (peak) late Arrhythmia/ECGqual NF ECG, skin (pattern) early Arrhythmia/APqual NF ECG, skin (pattern) early Arrhythmia/APqual NF ECG, skin (pattern) late Arrhythmia/ECGqual NF ECG, skin (pattern) early Arrhythmia/ECGqual NF ECG, skin (pattern) late Arrhythmia NF ECG, skin (peak) early ECG qual NF ECG, skin (pattern) late Arrhythmia NF ECG, skin (pattern) early ECG qual NF ECG, skin (pattern) late Arrhythmia FOS ECG, skin (pattern) late Arrhythmia FOS ECG, skin (pattern) late+early Arrhythmia/ECGqual NF ECG, skin (peak) late+early Arrhythmia FOS ECG, skin (peak) early Arrhythmia FOS ECG, skin (pattern) early Arrhythmia FOS ECG, skin (pattern) late Arrhythmia FOS ECG, skin (pattern) late Arrhythmia FOS ECG, skin (pattern) late Arrhythmia/ECGqual FOS ECG, skin (pattern) late Arrhythmia FOS ECG, skin (pattern) 0 APqual: poor quality arterial pressure signal, ECGqual: poor quality ECG signal, POS: position of the IAB catheter too low, Cath: catheter type, FOS: catheter with fibre-optic sensor, NF: catheter without fibre-optic sensor, ECG skin: ECG signal directly from patient skin, ECG MON: ECG signal slaved from monitor, AP: arterial pressure signal. Figure 1. Incidental incorrect timing with two inflations during systole (arrows) on strip nr 191, most likely caused by ECG artefacts. Table 4. Prevalence of poor quality ECG and arrhythmia during correct and incorrect timing Correct timing Poor quality ECG 11% 37% Arrhythmia 5% 69% Incorrect timing In our observations, we found poor quality pressure curves in 14 cases of all strips. Two of these cases presented with incorrect timing, but, in these cases, pattern trigger (using the ECG signal) was used. From our Figure 2. Arterial pressure signal with artefacts during diastolic augmentation. results, we do not conclude a poor quality blood pressure signal to be a cause of incorrect timing. With the use of a fibre-optic sensor, we did not find pressure curves of poor quality, although we did see artefacts during diastolic augmentation (Figure 2). These artefacts may result from whipping catheters or movement against the aortic wall. 14 Interpretation of inflation and deflation timing may be hampered by artefacts or poor quality pressure curves.

5 Bakker EWM et al. 5 Figure 3. Two cases of consistent early inflation (A=strip 71, B= strip 263). In five cases, we found a catheter presenting too low in the descending aorta on X-thorax. In three of these cases, timing was incorrect (consistent late deflation). After repositioning the catheter, timing was correct. Discussion Adding our observations of consistent and incidental incorrect timing, we found a total of 3% and 13% of inflations and deflations, respectively, to be incorrect. In 18 cases (5%), IABP assistance was switched to operator mode. We found consistently incorrect timing at some point of IABP assistance in 15 of 60 patients (25%). Our results should not be compared with the results of Osentowski, 10 using Datascope IABPs. We allowed a maximum of 40 ms deviation from the dicrotic notch or the end of diastole. Osentowski did not mention allowing any deviation from the dicrotic notch, and accepted no less than 5 to 10 mmhg reduction of end diastolic pressure. Furthermore, Osentowski only made a single evaluation per patient. According to Schreuder et al., 6 the fibre-optic sensor of the ultraflex catheter would effectively result in correct inflation timing. Inflation timing deviated only 0.6 ms from the dicrotic notch on average, with a range of 11 to +12 ms in 27 patients. We found four cases (with two patients) of consistent early inflation when using the fibre-optic sensor of the IABP (Figure 3). In these cases, switching to operator mode to adjust inflation timing would be justified. When discussing the accuracy of the autopilot during arrhythmias, early deflations could probably not be prevented if timing was assessed manually. The irregular rhythm would still cause the predictive nature of the algorithm to be disturbed, leading to timing discrepancies. Early deflations may be considered a safe way to handle irregular rhythms. The peak trigger, often used during arrhythmias, does not use R-wave deflation. It has a complex algorithm that continuously checks whether deflation empties the balloon for more than 40 percent before the start of the next systole. On the other hand, it determines its deflation point according to the average of a number of beats, so incidental late and early deflations will occur. It continuously adapts its deflation point to make it the safest, but still effective, point. Switching to the afib algorithm, which does use R-wave deflation, may prevent incidental timing errors during arrhythmias. The disadvantage of this trigger type is that all deflations may be relatively late. The cause of these late deflations could be the volumeregulated nature of the pneumatic system of the Autocat IABP, which may empty the balloon somewhat slower than a pressure-regulated pump. With the 2.23 software update, R-wave deflation can be provided without switching from autopilot mode to operator mode by turning arrhythmia timing on. In general, incorrect timing during arrhythmias may not always be prevented by switching to operator mode. Optimization of clinical care should, therefore, be focussed on resolving the irregular rhythm. One case of incidental incorrect timing is remarkable. We found two beats with inflations during systole (Figure 1). The landmarks show that the IABP misinterpreted the heavily distorted ECG signal. These incorrect inflations may have been prevented with a pacemaker trigger. The delay of the ECG signal when slaved from the monitor will influence timing. 15 We found four cases of monitor use in combination with incorrect timing, compared to seven cases with correct timing. The incidence is too low for conclusions from our observations. In one patient, we found late inflations on three strips (nr 34, 35 and 36 in Table 2) with very small systolic pressure curves followed by a strong augmentation curve. The duration of the systole was short (160 ms), as seen in hypovolemia. Later, during IABP assistance, the systolic time period increases to 240 ms, with a higher systolic pressure and correct IABP timing. In these cases of late inflations, the IABP used the pattern trigger mode, without fibre-optic sensor, resulting in a predictivepredictive timing method. If a fibre-optic sensor had been used, the wave-predictive algorithm might have prevented the late inflations. From our results, we recommend frequent checks of inflation and deflation timing. If incorrect timing is observed, switching to operator mode should be considered, particularly when timing is consistently incorrect. In a case of a poor ECG signal, a pacemaker trigger or

6 6 Perfusion 0(0) arterial pressure trigger may be considered if the signal remains poor. Timing errors during arrhythmias cannot always be prevented. The use of R-wave deflation (arrhythmia timing or afib trigger) to overrule the peak trigger may be considered if end diastolic pressure remains acceptable. Furthermore, a catheter with a fibre-optic sensor is recommended for its robust pressure signal and the application of the wave-predictive algorithm that may prevent late inflations. The clinical relevance of incorrect timing as seen in our observations has been a point of discussion during the last two decades. 16 Using high fidelity pressure measurements at the aortic root 17,18 or a conductance catheter to measure volume and pressure inside the left ventricle 9 may open the debate. We are aware of the fact that interpreting a print from a blood pressure measurement at the descending aorta is a method with low accuracy. Still, it is the clinical way of interpreting inflation and deflation timing in our daily clinical practice. Conclusions Inflation and deflation timing of the Arrow AutoCAT 2 WAVE, while using the autopilot mode, is not always correct. The quality of the ECG input signal and the occurrence of arrhythmia appear to be related to erroneous timing. Switching from autopilot mode to operator mode may not always prevent incorrect timing. Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. Conflict of Interest Statement None declared. References 1. Trost JC, Hillis D. Intra-aortic balloon counterpulsation. Am J Cardiol 2006; 97: Kantrowitz A, Kantrowitz A. Experimental augmentation of coronary flow by retardation of the arterial pressure pulse. Surgery 1953; 34: Zelano JA, LI JKJ, Welkowitz W. A closed loop control scheme for intraaortic balloon pumping. IE EE Trans Biomed Eng 1990; 37: Schreuder JJ, Maisano F, Donelli A, et al. Beat-to-beat effects of intraaortic balloon pump timing on left ventricular performance in patients with low ejection fraction. Ann Thorac Surg 2005; 79: Reesink KD, Nagel T, Bovelander J, et al. Feasibility study of a fiber-optic system for invasive blood pressure measurements. Cathet Cardiovasc Intervent 2002; 57: Schreuder JJ, Castiglioni A, Donelli A, et al. Automatic intraaortic balloon pump timing using an intrabeat dicrotic notch prediction algorithm. Ann Thorac Surg 2005; 79: Wesseling KH, Jansen JRC, Settels JJ, Schreuder JJ. Computation of aortic flow from pressure in humans using a nonlinear, three-element model. J Appl Physiol 1993; 74: Donelli A, Jansen JRC, Hoeksel B, et al. Performance of a real-time dicrotic notch detection and prediction algorithm in arrhythmic human aortic pressure signals. J Clin Monit 2002; 17: Hoeksel SAAP, Jansen JRC, Blom JA, Schreuder JJ. Detection of dicrotic notch in arterial pressure. J Clin Monit 1997; 13: Osentowski MK, Holt DW. Evaluating the efficacy of intra-aortic balloon pump timing using the auto-timing mode of operation with the Datascope CS100. J Extra Corpor Technol 2007; 39: Patel SI, Souter MJ. Equipment-related electrocardiographic artifacts: causes, characteristics, consequences, and correction. Anesthesiology 2008; 108: Sakiewicz PG, Wright E, Robinson O, et al. Abnormal electrical stimulus of an intra-aortic balloon pump with concurrent support with continuous veno-venous hemodialysis. ASAIO J 2000; 46: Bernstein AD, Parsonnet V. R-wave triggering of external instrumentation during cardiac pacing. PACE 1983; 6: Parissis H, Leotsinidis M, Dougenis D, Richens D. The way the intra-aortic balloon catheter moves within the aorta as a possible mechanism of balloon associated morbidity. Interact CardioVasc Thorac Surg 2007; 6: Arrow International. Operatoring Manual, Arrow Autocat tm 2 Series Intra-Aortic Balloon Pump (IABP) System, IAM-9005, Rev. 1, 5.1 ECG connections, Kern MJ, Aquirre FV, Caracciolo EA, et al. Hemodynamic effects of new intra-aortic balloon counterpulsation timing methods in patients: a multicenter evaluation. Am Heart J 1999; 137: Pantalos GM, Koenig SC, Gillars KJ, Haugh GS, Dowling RD, Gray LA Jr. Intraaortic balloon pump timing discrepancies in adult patients. Artif Organs 2011; 35: Pantalos GM, Minich LL, Tani LY, McGough EC, Hawkins JA. Estimation of timing errors for the intraaortic balloon pump use in pediatric patients. ASAIO J 1999; 45: