Continuous-flow (CF) left ventricular assist devices. Original Article

Transcription

1 Original Article Validity and Reliability of a Novel Slow Cuff-Deflation System for Noninvasive Blood Pressure Monitoring in Patients With Continuous-Flow Left Ventricular Assist Device Gregg M. Lanier, MD*; Khristine Orlanes, NP*; Yacki Hayashi; Jennifer Murphy, NP; Margaret Flannery, NP; Rosie Te-Frey, NP; Nir Uriel, MD; Melana Yuzefpolskaya, MD; Donna M. Mancini, MD; Yoshifumi Naka, MD; Hiroo Takayama, MD; Ulrich P. Jorde, MD; Ryan T. Demmer, PhD*; Paolo C. Colombo, MD* Background Doppler ultrasound is the clinical gold standard for noninvasive blood pressure (BP) measurement among continuous-flow left ventricular assist device patients. The relationship of Doppler BP to systolic BP () and mean arterial pressure () is uncertain and Doppler measurements require a clinic visit. We studied the relationship between Doppler BP and both arterial-line (A-line) and. Validity and reliability of the Terumo BP Monitor, a novel slow cuff-deflation device that could potentially be used by patients at home, were assessed. Methods and Results Doppler and Terumo BP measurements were made in triplicate among 6 axial continuousflow left ventricular assist device (HeartMate II) patients (3 inpatients and 3 outpatients) at 2 separate exams (36 possible measurements). A-line measures were also obtained among inpatients. Mean absolute differences (MADs) and correlations were used to determine within-device reliability (comparison of second and third BP measures) and betweendevice validity. Bland Altman plots assessed BP agreement between A-line, Doppler BP, and Terumo. Success rates for Doppler and Terumo were 1% and 91%. Terumo MAD for repeat and were 4.6±.6 and 4.2±.6 mm Hg; repeat Doppler BP MAD was 2.9±.2 mm Hg. Mean Doppler BP was lower than A-line by 4.1 (MAD=6.4±1.4) mm Hg and higher than by 9. (MAD=11.±1.2) mm Hg; Terumo underestimated A-line by.3 (MAD=.6±.9) mm Hg and by 1.7 (MAD=6.±1.) mm Hg. Conclusions Doppler BP more closely approximates than. Terumo was successful, reliable, and valid when compared with A-line and Doppler. (Circ Heart Fail. 213;6:1-112.) Key Words: blood pressure measurement/monitoring heart-assist devices heart failure Terumo Continuous-flow (CF) left ventricular assist devices (LVADs) have become the standard of care in patients with end-stage heart failure as bridge to transplantation and destination therapy. They offer several advantages compared with the older version pulsatile devices, such as smaller size, enhanced durability, and better long-term outcomes. 1,2 Editorial see p 879 Clinical Perspective on p 112 Accurate monitoring of arterial blood pressure (BP) is required for optimization of early postoperative and of longterm care in CF-LVAD patients. An arterial line (A-line) is routinely used to measure BP during the initial postoperative period. Once the A-line is removed, BP can be obtained noninvasively using automated BP monitors or Doppler ultrasound. Automated BP monitors are accurate, but their success rate is low ( %), 3 due to reduced pulse pressure (PP) in CF-LVAD patients. Consequently, Doppler ultrasound is commonly used to assess BP in patients with CF-LVAD during clinical visits. Current clinical understanding is that Doppler BP represents mean arterial pressure () rather than systolic BP (); however this remains controversial. 3 Although Doppler assessments have excellent success rates ( 9%) and high validity, 3 these measurements are timeconsuming and cannot be performed by patients at home, requiring highly trained personal in the hospital setting. In an attempt to overcome the shortcomings of current noninvasive methods for BP measurement in CF-LVAD patients, Received November 2, 212; accepted June 2, 213. From the New York Presbyterian, Columbia University, New York, NY (G.M.L., K.O., Y.H., J.M., M.F., R.T.-F., N.U., M.Y., D.M.M., Y.N., H.T., U.P.J., P.C.C.); Westchester Medical Center, Valhalla, NY (G.M.L.); and Department of Epidemiology, Mailman School of Public Health, Columbia University, New York, NY (R.T.D.). *Drs Lanier, Orlanes, Demmer, and Colombo contributed equally to this article. The online-only Data Supplement is available at /-/DC1. Correspondence to Paolo C. Colombo, MD, New York Presbyterian, Columbia University, 622 W 168th St #Ph12-134, New York, NY pcc21@columbia.edu 213 American Heart Association, Inc. Circ Heart Fail is available at DOI: /CIRCHEARTFAILURE

2 16 Circ Heart Fail September 213 we tested the success rate, reliability, and validity of a novel slow cuff-deflation system, which enhances sensitivity for detection of BP in patients with reduced arterial PP. The Terumo BP Monitor was compared with BP measurements taken with either A-line (gold standard) or Doppler ultrasound (clinical standard). We also compared the performance of an automated BP monitor (GE CARESCAPE V1) with A-line and Doppler, as this automated device is routinely used at our Institution, in patients on CF-LVAD support. Methods Patient Population A total of 6 patients were prospectively studied at 2 time points as follows: (1) 3 inpatients were assessed immediately following implantation of an axial flow CF-LVAD, the Heart Mate II (Thoratec Corporation, CA; inpatient visit v1), and the second encounter (inpatient v2) occurred a minimum of 3 days later after the patient was no longer in the intensive care unit and no longer had an A-line; and (2) a separate group of 3 outpatients on Heart Mate II LVAD support, who presented to clinic for routine follow-up, were assessed at 2 separate visits (outpatient v1 and outpatient v2, respectively). Patients enrolled in the inpatient group could not be enrolled in the outpatient group. The study protocol was approved by the local institutional review board and all subjects signed provided informed consent. Study Protocol During their first encounter (inpatient v1), LVAD inpatients had BP measurements by a radial A-line obtained immediately prior to each of the 3 noninvasive modalities tested (see below). Before measurements were recorded, the A-line was flushed once and leveled to the pressure transducer (IntelliVue, Phillips, the Netherlands). BP measurements were done in triplicate for each method. A-line BP was recorded in a paired fashion, prior to each of the triplicate measurements by each noninvasive modality. All BP measurements were obtained in the same arm as the A-line. For all patients, 3 separate BPs were taken 4 with each of the following 3 different noninvasive BP measurement methods: Doppler ultrasound (Lumeon Doppler System, Houston, TX), GE CARESCAPE V1 (Fairfield, CT) automated BP monitor, and Terumo BP (Somerset, NJ) monitor. Doppler ultrasound BP measurements were obtained using a calibrated sphygmomanometer as previously described. 4 Briefly, the Doppler probe was placed over the brachial artery while the examiner verified that the brachial pulse could be auscultated. The cuff was placed proximal to the Doppler probe, attached to a manometer, and inflated until the pulse was no longer audible by Doppler. The cuff was then slowly deflated (2 3 mm Hg/s) allowing the reestablishment of blood flow and a reading was taken when the pulse became audible again. The Terumo BP monitor uses a double-cuff oscillometric, slow-deflation technology. BP measures were taken according to manufacturer defaults. The GE CARESCAPE V1 monitor was chosen for its routine use at our hospital, to be taken as representative of a standard automated BP monitor. Appropriately sized BP cuffs were chosen for all noninvasive modalities, as previously described. 4 BP measurements were taken in the upper arm in triplicate, and the average of the last 2 measurements was used for subsequent analyses. 4 Because the Terumo provides only the, diastolic BP, and heart rate (HR), was calculated as follows: (+2 diastolic BP)/3. The GE CARESCAPE V1 monitor provides a automatically. Whenever possible, the same settings were used during the first and second encounter: the patient was seated, with legs uncrossed, and the back and arm supported, such that the middle of the cuff on the upper arm is at the level of the right atrium (the midpoint of the sternum). In the CTICU, the patient was permitted to be supine. All noninvasive measurements were taken in the same arm as the A-line. BP measures were performed in the following order: Doppler ultrasound, GE CARESCAPE V1, and Terumo. The second encounter (inpatient v2) occurred a minimum of 3 days later and when the patient was no longer in the intensive care unit and no longer had an A-line. The BP measurements for outpatients were performed on 2 separate follow-up visits (outpatient v1 and outpatient v2), and with the same methodology and sequence described above, except that no A-line measurement was available. Ability to provide a reading on any BP measurement attempt was defined as success. At the time of each encounter, HR and Heart Mate II LVAD data including LVAD speed and pulsatility index were recorded. In inpatients, HR regularity, that is, regular versus irregular, was also assessed, based on ECG and telemetry monitoring. Of note, patients in atrial fibrillation but with regular ventricular pacing were labeled as regular; atrial fibrillation with an irregular ventricular rate and sinus rhythm with frequent premature atrial or ventricular complexes were labeled as irregular. Two inpatients had their A-line data recorded at the beginning of the first visit, that is, prior to Doppler, but not prior to subsequent BP measurements using the other modalities; in these 2 patients these initial (and only) A-line readings were used for all comparisons. Additionally, 2 inpatients did not have a second set of BP measurements during a second visit (v2) prior to hospital discharge because of logistical reasons. Statistical Analysis All analyses were performed in SAS version 9.3. Descriptive data are presented as proportions or means±sd (unless otherwise specified). Reliability and validity were assessed using multiple approaches. For reliability, mean absolute differences (MAD) were computed between the second and third repeat measures within each device. When MADs are presented for the overall patient sample which includes repeat BP measurements for each patient at 2 time points a no-intercept random-effects mixed regression model with patient modeled as a random effect was used to generate SEs accounting for the correlated nature of the data. The default correlation structure for these models is variance components. Intraclass correlation coefficients were also calculated using all 3 BP measures among subgroups defined by patient (inpatient or outpatient) and time (v1 or v2). To assess validity, MADs were calculated comparing BP values derived from Doppler, Terumo, and GE CARESCAPE V1 versus A-line among n=3 inpatients. Bland Altman plots were constructed for various comparisons of agreement between Terumo, Doppler, and A-line systolic BP measurements. Because 2 within-patient measures were averaged to compute mean BP values, a corrected SD was computed as previously described. 6 Similarly, among outpatients, MADs were calculated comparing BP values derived from Terumo and GE CARESCAPE V1 versus Doppler, and a Bland Altman plot was constructed comparing Terumo and Doppler BP. Mean between-device differences for Bland Altman plots are presented (as compared with MAD) because this provides information on whether or not comparison measurements over- or underestimate gold standard measures. Therefore, we present both MADs and mean differences for between-device comparisons as both measures provide unique and important information. As a secondary analysis, we also present standard Pearson correlation coefficients as a crude metric of both within- and between-device agreement because they are commonly understood statistics in the medical literature; however, the reader should be cautioned that they are not ideal for studies of reliability and validity 6 as Pearson correlation coefficients estimate degree of linear association between variables and not necessarily degree of similarity. When assessing associations between clinical parameters and device measurement success, we used generalized linear models (PROC GENMOD) in SAS with a logit link to generate odds ratios summarizing these associations. This method allowed us to the multiple within-patient measurements taken (as opposed to 1 measure per patient) and properly account for the correlation of outcomes within patient. Finally, ANOVA was used to compare mean PP among patients with 3 successful BP readings versus <3 successful BP readings.

3 Lanier et al Blood Pressure Monitoring in CF-LVADs 17 Results Patient Characteristics Baseline characteristics of patients are listed in Table 1. Overall, the average age of patients was 6 years, 7% had heart failure of ischemic etiology, 7% had a Heart Mate II LVAD implanted as a bridge to transplantation. The first inpatient visit (v1) occurred, on average, 4±2 (range, 1 9) days, postoperatively and the second visit (v2) occurred an average of 13± (range, 2 24) days after the first encounter. Outpatient visits (v1 and v2) were separated by an average of 49±3 (range, 8 13) days. One notable difference between inpatients and outpatients was the HR, which was higher among inpatients. Irregularity of HR was noted in 21% of inpatients. Measurement Success Among 6 patients (8 of which had both per-protocol encounters) the total number of BP measurement attempts made for Doppler, Terumo, and GE CARESCAPE V1 were 34. The overall success rate for Doppler was 1%. For Terumo and GE CARESCAPE V1, the success rates were 91% and 63%, respectively (Table 2). A-line PP (range, 3 41 mm Hg; median, 2 mm Hg) and HR regularity were associated with Terumo and GE CARESCAPE V1 measurement success. Among inpatients with A-line measurements, a 9 mm Hg (equivalent to 1 SD) increase in PP was associated with an 1% increase in the odds of Terumo success (P=.6) and an 3% increase in the odds of GE CARESCAPE V1 success (P<.1); respective odds ratios [9% confidence interval] for Terumo and GE CARESCAPE V1 success are 1.99 [.98, 2.6] and 3.93 [1.9, 7.81]. The mean A-line PP with <3 successful Terumo readings versus those with 3 successful Terumo readings was 1 versus 22 mm Hg (P=.6). Similarly, mean A-line PP was 1 versus 23 mm Hg (P=.9) among patients with <3 versus 3 successful GE CARESCAPE V1 readings (P=.33). Cardiac rhythm was also associated with measurement success. A regular HR was strongly associated with an increase in the odds of GE CARESCAPE V1 success; 7. [2.3, 21.], P<.1. However, results were substantially weaker and nonstatistically significant for associations between Table 1. Characteristics of Patients With Continuous-Flow Left Ventricular Assist Device Inpatients (n=3) Outpatients (n=3) Age, y 6±13 9±13 Male, % 9 76 Ischemic etiology, % 3 6 Bridge to transplant, % 7 7 LVAD speed, rpm v1: 8831±343 v2: 8811±28 Pulsatility index v1:.±.7 v2:.2±.9 Heart rate, bpm v1: 92±14 v2: 87±9 v1: 8882±46 v2: 8886±42 v1:.3±.9 v2:.4±.8 v1: 68±11 v2: 69±11 v1 represents the first patient encounter and v2 represents the second patient encounter. LVAD indicates left ventricular assist device. regular cardiac rhythm and Terumo success: 2.1 [., 9.3], P=.33. Reliability of Repeat Within-Device Measures All devices demonstrated good reliability. The overall correlation between the second and third measures for A-line, Table 2. Blood Pressure Measurement Success Rates [9% Confidence Intervals] by Device Among Patients With Continuous-Flow Left Ventricular Assist Device Percent of Total Percent of Patients Blood Pressure With 3 Successful Measurement Blood Pressure Attempts That Were Measurement Successful Attempts Percent of Patients With Successful Second and Third Blood Pressure Measurement Attempts Inpatient visit 1* A-line Doppler Terumo 88 [78, 94] 73 [6, 89] 77 [62, 92] GE CARESCAPE 71 [7, 82] 7 [39, 74] 63 [46, 81] V1 Inpatient visit 2* Doppler Terumo 83 [67, 93] 8 [66, 94] 8 [66, 94] GE CARESCAPE 7 [42, 7] 37 [19, 4] [32, 68] V1 Outpatient visit 1* Doppler Terumo 92 [8, 96] 77 [62, 92] 9 [79, 1] GE CARESCAPE 7 [43, 7] 33 [16, ] 47 [29, 6] V1 Outpatient visit 2* Doppler Terumo 96 [89, 98] 87 [7, 99] 97 [9, 1] GE CARESCAPE V1 61 [46, 74] 43 [26, 61] 3 [3, 71] Combined inpatient and outpatient visit 1 and visit 2 Doppler Terumo 9 [8, 93] 79 [72, 86] 86 [8, 92] GE CARESCAPE V1 61 [3, 69] 43 [37, 1] 3 [44, 62] *The denominator for percent of total blood pressure measurement attempts that were successful for any particular clinical setting and time point (eg, inpatient visit1) was 9. The denominator was 3 for percent of patients with 3 successful blood pressure measurement attempts, and for percent of patients with successful second and third blood pressure measurement attempts for any particular clinical setting and time point, it was 3. Two inpatients were not examined at visit 2 for logistical reasons (see Methods section), yielding a total number of 28 patients and 84 blood pressure measurement attempts among inpatients at visit 2. The denominator when combining all clinical settings and time points was 118 patients (118 instead of 12 because 2 patients were not examined at visit 2) and 34 blood pressure measurement attempts (34 vs 36 attempts because 2 inpatients were not examined at v2).

4 18 Circ Heart Fail September 213 Doppler, Terumo, and GE CARESCAPE V1 were:.99 (P<.1),.9 (P<.1),.82 (P<.1),.93 (P<.1). For, the overall correlation between the second and third measures for A-line, Terumo, and GE CARESCAPE V1 were:.98 (P<.1),.87 (P<.1),.9 (P<.1). Table I in the online-only Data Supplement presents inter- and intraclass correlation coefficients for repeat measures according to device, clinical setting, and time point. Overall, the respective MADs±SE between second and third measures for A-line, Doppler, Terumo, and GE CARESCAPE V1 were: 1.6±.3, 2.9±.2, 4.6±.6, and 3.7±. mm Hg. For, the respective MADs±SE between second and third measures for A-line, Terumo, and GE CARESCAPE V1 were: 1.2±.4, 4.2±.7, and 2.7±.3 mm Hg. Within-device MAD values for, diastolic BP, and across all methods, clinical setting, and patient visit are presented in Table II in the online-only Data Supplement. Validity of Doppler, Terumo, and GE CARESCAPE Versus A-line (Gold Standard) Doppler, Terumo, and GE CARESCAPE V1 measurements demonstrated strong correlations (ranging from.73 to.88; all P<.1) with gold standard A-line measurements (Table 3). Overall, the respective MADs±SE between A-line and either Doppler BP, Terumo, or GE CARESCAPE V1 were: 6.4±1.4,.6±.9, and.2±.7 mm Hg. While successful in fewer patients, GE CARESCAPE V1 often performed as well as Doppler and Terumo, and in some instances marginally better, depending on the clinical setting (Table 3). As shown in Table 3, restricting the analysis to a fixed sample of patients on which all devices yielded a measurement (n=17) had minimal influence on which device performed best. Among both inpatients and outpatients, Doppler BP tended to more closely reflect rather than by A-line, Terumo, or GE CARESCAPE V1 (Figure 1). For example, among inpatients, the mean differences (as opposed to MAD) between Doppler BP and A-line BP showed that Doppler BP was lower than by 4.1 mm Hg, while Doppler BP was higher than by 9. mm Hg. However, although Doppler BP consistently reflects, in situations of low PP Doppler BP may also closely approximate. For example, the mean difference between Doppler BP and A-line was consistent among patients with A-line PP below the median (2 mm Hg) versus above the median:. versus 3.2 mm Hg, respectively (P=.46). In contrast, the mean difference between Doppler BP and A-line was markedly different among patients with A-line PP below the median versus above the median: 3. versus 13.6 mm Hg, respectively (P<.1). Similarly, Figure 2 shows that increased A-line PP is also related to a greater difference between Doppler BP and Terumo. Nevertheless, despite a strong correlation, there remains a sizeable number of patients who have a difference between Doppler BP and Terumo of > mm Hg even when PP is low. This observation reinforces the Table 3. Between-Device Mean Absolute Differences (SE) and Correlations Among Patients With Continuous-Flow Left Ventricular Assist Device Method Inpatient visit 1, A-Line vs Doppler (n=3) Doppler (n=17)* Terumo (n=23) Terumo (n=17)* GE CARESCAPE V1 (n=19) GE CARESCAPE V1 (n=17)* Inpatient visit 2, Doppler vs Terumo (n=24) Terumo (n=14)* GE CARESCAPE V1 (n=1) GE CARESCAPE V1 (n=14)* Outpatient visit 1, Doppler vs Terumo (n=27) Terumo (n=12)* GE CARESCAPE V1 (n=14) GE CARESCAPE V1 (n=12)* Outpatient visit 2, Doppler vs Terumo (n=29) Terumo (n=16)* GE CARESCAPE V1 (n=16) GE CARESCAPE V1 (n=16)* Systolic Blood Pressure Diastolic Blood Pressure Mean Arterial Pressure MAD (SE) r (P Value) MAD (SE) r (P Value) MAD (SE) r (P Value) 6.4 (1.4) 6.3 (1.4).6 (.9).4 (1.2).2 (.7).1 (.8).2 (1.).6 (1.4) 7.1 (1.9).4 (1.).7 (.8) 3.8 (1.) 4.3 (.9) 4.1 (1.).7 (1.) 6. (1.3).7 (1.).7 (1.).79 (<.1).73 (.9).83 (<.1).78 (.2).88 (<.1).86 (<.1).91 (<.1).9 (<.1).81 (<.1).93 (<.1).88 (<.1).96 (<.1).9 (<.1).92 (<.1).78 (<.1).78 (.3).8 (.2).8 (.2) 17.9 (1.7) 22.3 (2.) 7.4 (1.3) 7.8 (1.3).3 (1.2).7 (1.3) 1.3 (2.1) 18.2 (2.9) 2.2 (2.3) 2.4 (2.) 16.3 (1.7) 2. (2.9) 26. (2.1) 2 (2.2) 17.9 (1.7) 19.2 (2.) 27.1 (1.7) 27.1 (1.6).44 (.1).34 (.18).7 (.4).6 (.1).6 (.3).62 (.8).74 (<.1).77 (.1).84 (<.1).84 (.2).61 (.8).6 (.6).7 (.2).76 (.4).8 (.1).63 (.9).8 (.2).8 (.2) 11. (1.2) 12.9 (1.) 6. (1.) 6.4 (1.1) 4.3 (1.) 4.6 (1.) 9. (1.6) 11.4 (2.4) 1. (2.) 16. (1.9) 9.9 (1.3) 12.3 (2.) 16.9 (1.7) 1.7 (1.7) 12.1 (1.4) 12.8 (2.1) 18.2 (1.) 18.1 (1.).64 (<.1).3 (.3).7 (<.1).69 (.2).86 (<.1).83 (<.1).8 (<.1).8 (.1).88 (<.1).91 (<.1).7 (<.1).81 (.1).84 (.2).87 (.3).68 (<.1).7 (.3).84 (<.1).84 (<.1) Comparisons restricted to patients with second and third measures for each respective device comparison, clinical setting (inpatient vs outpatient), and time point (visit 1 vs visit 2). MAD indicates mean absolute difference. *Restricted to sample where all patients have data for all comparisons to standardize the sample among which reliability is compared.

5 Lanier et al Blood Pressure Monitoring in CF-LVADs 19 A Blood Pressure (mmhg) B Blood Pressure (mmhg) Doppler A-line Terumo GE CARESCAPE V1 Doppler Terumo GE CARESCAPE V1 Figure 1. Mean blood pressure values (±SE) among: (A) inpatients according to measurement device among 23 inpatients (results for GE CARESCAPE V1 include n=17 patients with successful measures); and (B) outpatients according to measurement device among 27 outpatients (results for GE CARESCAPE V1 include n=12 patients with successful measures). indicates diastolic blood pressure;, mean arterial pressure; and, systolic blood pressure. importance of using Terumo rather than Doppler to assess in CF-LVAD patients. Bland Altman plots demonstrate that agreement between Terumo and A-line ( or ) did not vary according to BP level. Ninety-five percent of Terumo measurements were within 1 mm Hg of A-line (Figure 3A) and 9% of Terumo measurements were within 16 mm Hg of A-line (Figure 3B). The Bland Altman plot was similar for Doppler versus A-line (Figure 3C). Validity of Terumo and CARESCAPE V1 Versus Doppler (Clinical Standard) Among all patients (at both visits) with successful second and third BP readings, the overall correlations between mean Doppler BP and mean assessed using either Terumo or GE CARESCAPE V1 were.8 (P<.1; Figure 4A) and.81 (P<.1). The overall Mean Difference between Doppler and Terumo Mean Arterial Blood Pressure (mm Hg) r=.48, p= Arterial-Line Pulse Pressure (mm Hg) MADs between Doppler BP and mean assessed using either Terumo or GE CARESCAPE V1 were.7±.7 and 7.1±1.2 mm Hg. Table 3 shows the betweendevice MADs and correlation coefficients comparing Doppler BP with Terumo and GE CARESCAPE V1 among patient subgroups according to clinical setting and time point. The Bland Altman plot for Doppler versus Terumo is shown in Figure 4B. Of note, pump speed was not associated with agreement between Doppler BP and Terumo ; for a 1 point decrease in pump speed, there was <1 point, nonstatistically significant increase in the MAD between Doppler BP and Terumo (P=.3). Discussion This study is the largest published series to date analyzing invasive and noninvasive BP measurements in axial CF-LVADs, in both the inpatient and outpatient setting. The results indicate Figure 2. Scatterplot of the difference between Doppler blood pressure and Terumo mean arterial pressure plotted against arterial-line pulse pressure.

6 11 Circ Heart Fail September 213 A Difference between Terumo and Arterial-Line Systolic Blood Pressure (mmhg) Mean difference = -.34 mmhg Corrected STD = Average of Terumo and Arterial-Line Systolic Blood Pressure (mmhg) that (1) Doppler BP better reflects rather than ; (2) Terumo BP monitor has a high rate of measurement success, particularly in comparison to traditional automated BP devices; and (3) Terumo provides an accurate and reliable measurement of and. Throughout the past years axial and more recently centrifugal CF-LVADs have replaced pulsatile devices (as bridge to transplantation or destination therapy) due to better outcomes and fewer complications in end-stage heart failure patients. Regular and reliable assessment of arterial BP is important for optimization of cardiovascular care postoperatively and after discharge from the hospital. In the early postoperative period, an A-line is routinely used for BP monitoring. After the A-line is removed, noninvasive modalities of assessment are critical for adequate patient care. However, due to the reduced PP, which may fall below the sensitivity range of auscultatory and automated BP methods, it is often difficult or impossible to measure BP using standard techniques during CF-LVAD support. Consequently, current recommendations are to use Doppler ultrasound for arterial BP measurement in patients. 7 Clinical recommendations are to maintain the in the range of 7 to 8 mm Hg and not to exceed 9 mm Hg, as high afterload my compromise unloading of the left ventricle. 7 C Difference between Doppler and Arterial-Line Systolic Blood Pressure (mmhg) B Difference between Terumo & Arterial-Line Mean Arterial Pressure (mmhg) Mean difference = -4.1 mmhg Corrected STD = Mean difference = -1.7 mmhg Corrected STD = Average of Terumo and A-Line Mean Arterial Pressure (mmhg) Average of Doppler and A-Line Systolic Blood Pressure (mmhg) Figure 3. Bland Altman plots comparing: (A) A-line systolic blood pressure to Terumo systolic blood pressure, mean difference=.34, corrected SD=7.9; (B) A-line mean arterial pressure to Terumo mean arterial pressure, mean difference= 1.7, corrected SD=8.3; (C) A-line systolic blood pressure to Doppler blood pressure, mean difference= 4.1, corrected SD=7.4. Mean blood pressure differences were calculated as Terumo minus A-line (Figure 2A and 2B) or Doppler minus A-line (Figure 2C). Although there is general agreement on these recommendations, there is less consensus on how to accurately measure BP in patients on support with a CF-LVAD. Additionally, without having direct A-line measurements, it is unclear whether Doppler BP represents the or in the absence of a normal arterial waveform, although, based on limited published data, 3 many care providers have considered the Doppler BP to reflect the. A Doppler ultrasound probe can detect flow at any point during the cardiac cycle and can measure a pressure in almost all CF-LVAD patients as demonstrated in our data and by others. 3 Knowing whether Doppler more closely measures or is important for clinical decision making, but few studies provide data to address this issue. One group reported that in 17 patients who underwent HeartMate II implantation, the mean difference between Doppler and A-line was.2±1. mm Hg compared with a mean difference between Doppler and A-line of 8.6±9. mm Hg, thus suggesting that Doppler BP values more closely represent the true as opposed to true. 3 Conversely, we found that Doppler BP underestimates by 4.1±1. mm Hg and overestimates by 9.±1.9 mm Hg, thus suggesting that Doppler better reflects.

7 Lanier et al Blood Pressure Monitoring in CF-LVADs 111 A 1 Terumo Systolic Blood Pressure (mmhg) r=.8, p< Doppler Blood Pressure (mmhg) B Difference between Doppler and Terumo Systolic Blood Pressure (mmhg) Mean difference = -3.1 mmhg Corrected STD = Average of Terumo and Doppler Systolic Blood Pressure (mmhg) Figure 4. A, Scatterplot of mean Doppler systolic blood pressure versus mean Terumo systolic blood pressure among inpatients and outpatients at both visits 1 and 2; and (B) Bland Altman plot comparing Doppler blood pressure to Terumo systolic blood pressure among inpatients and outpatients at visit 1; mean difference= 3.1, corrected SD=7.3. Only visit 1 results presented so that the appropriate corrected SD could be presented. Results from visit 2 were consistent with those from visit 1. We further observed that the relationship between Doppler BP and was not meaningfully affected by PP, while the relationship between Doppler BP and was substantially influenced by PP. This is expected from a hemodynamic perspective because Doppler detects arterial blood flow, which will consistently occur as soon as the cuff pressure falls below the. Therefore, lower PPs in the previous study might have narrowed the difference between and resulting in Doppler BP values that were closer to the values. Other methodological differences in acquisition of Doppler BP, that is, slower versus faster rate of cuff release, may also explain the discrepancy between our study and others. 3 This observation has important clinical consequences, as it suggests that titration of outpatient vasoactive medications to a Doppler BP of 7 to 8 mm Hg, as has been suggested previously, may produce adverse side effects of hypotension, such as dizziness or syncope, among patients with substantial pulsatililty where Doppler is a better reflection of than. From this perspective, our data suggest that Terumo BP Monitor offers a more useful tool to meet recommended goals not only in the outpatient, but also in the inpatient setting when A-line measurements are no longer available. Other critical limitations of Doppler methodology are its time-consuming nature and requirement for trained personal in the hospital setting. The Terumo BP Monitor has a slow-deflation setting, which is designed to improve the sensitivity of BP measurements in patients with a reduced PP. In this study, the Terumo BP Monitor was successful, reproducible, and valid when compared with A-line (gold standard) and Doppler (clinical standard). Specifically, we show that Terumo measures have less bias in estimating A-line (Figure 3A; Terumo versus A-line) than Doppler (Figure 3C) and importantly that Terumo provides a reasonable estimate of in LVAD patients. This portable BP device is relatively inexpensive ( $2), weighs only12 g, and is powered by a single AA battery. While most centers currently use Doppler ultrasound for BP management, our results suggest that the Terumo BP monitor, when successful, provides an accurate and more applicable alternative. In addition, it also offers other advantages, which are typical of automated BP devices, such as the elimination of observer error and white coat effect, and the potential for more BP readings. A standard automated BP monitor, the GE Dinamap CARESCAPE V1, was also tested in this study. Our results indicate that this method has a success rate of only 63% compared with 91% of Terumo, a result which is consistent with the literature on standard automated BP methodologies. 3 Nevertheless, when obtainable, readings are accurate and can be used to direct patient care. However, the size (>2 kg) of this on-wheel device and its cost (>$2) make the use of this automated BP monitor impractical, particularly in the outpatient setting. Limitations of our study include the lack of blinding between the A-line recordings and the Doppler ultrasound recordings. This may have led to unintentional bias toward (or against) Doppler BP agreement. However, our Doppler correlation results seem consistent overall with previous publications. 3 Also, that Terumo BP measurement is overall successful and valid does not mean that it will provide accurate readings in all patients. There were a substantial number of study subjects in whom the measurement difference compared with A-line (gold standard) or Doppler (current clinical standard) was >1 mm Hg. For this reason, we agree with the current American Heart Association recommendation that Terumo, as any oscillometric BP monitor, should be validated on each patient, that is, against A-line and Doppler postoperatively after LVAD implantation, before the readings are accepted as reliable and valid. 4 One additional limitation is that our study did not include patients supported with centrifugal CF-LVADs. The present study shows that (1) the Terumo BP monitor offers reliable measurement of BP in the majority of patients on axial CF-LVAD support; (2) it correlates closely

8 112 Circ Heart Fail September 213 with and adequately with by A-line; (3) the difference between and varies between patients, such that (4) the target for BP management remains to be established for patients with substantial pulsatility. Based on these conclusions, when the targets for BP have been identified for a given patient, the Terumo offers potential for valid and reliable long-term management using home BP monitoring. Sources of Funding This work was supported by Thoratec Corporation from an educational grant to Columbia University Medical Center. Disclosures Drs Jorde and Naka are consultants to Thoratec Corporation. The other authors report no conflict. References 1. Slaughter MS, Rogers JG, Milano CA, Russell SD, Conte JV, Feldman D, Sun B, Tatooles AJ, Delgado RM III, Long JW, Wozniak TC, Ghumman W, Farrar DJ, Frazier OH; HeartMate II Investigators. Advanced heart failure treated with continuous-flow left ventricular assist device. N Engl J Med. 29;361: Miller LW, Pagani FD, Russell SD, John R, Boyle AJ, Aaronson KD, Conte JV, Naka Y, Mancini D, Delgado RM, MacGillivray TE, Farrar DJ, Frazier OH; HeartMate II Clinical Investigators. Use of a continuousflow device in patients awaiting heart transplantation. N Engl J Med. 27;37: Bennett MK, Roberts CA, Dordunoo D, Shah A, Russell SD. Ideal methodology to assess systemic blood pressure in patients with continuous-flow left ventricular assist devices. J Heart Lung Transplant. 21;29: Pickering TG, Hall JE, Appel LJ, Falkner BE, Graves J, Hill MN, Jones DW, Kurtz T, Sheps SG, Roccella EJ. Recommendations for blood pressure measurement in humans and experimental animals: part 1: blood pressure measurement in humans: a statement for professionals from the Subcommittee of Professional and Public Education of the American Heart Association Council on High Blood Pressure Research. Circulation. 2;111: Tochikubo O, Nishijima K, Ohshige K, Kimura K. Accuracy and applicability of the Terumo ES-H double-cuff sphygmomanometer for hospital use. Blood Press Monit. 23;8: Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1: Slaughter MS, Pagani FD, Rogers JG, Miller LW, Sun B, Russell SD, Starling RC, Chen L, Boyle AJ, Chillcott S, Adamson RM, Blood MS, Camacho MT, Idrissi KA, Petty M, Sobieski M, Wright S, Myers TJ, Farrar DJ; HeartMate II Clinical Investigators. Clinical management of continuous-flow left ventricular assist devices in advanced heart failure. J Heart Lung Transplant. 21;29(suppl 4):S1 39. CLINICAL PERSPECTIVE This study prospectively analyzes invasive and noninvasive blood pressure (BP) measurements in patients on continuousflow left ventricular assist device support. Due to the reduced pulse pressure, noninvasive BP measurements are challenging in continuous-flow left ventricular assist device patients; hence, Doppler ultrasound is frequently used. However, the relationship of Doppler BP to systolic BP and mean arterial pressure remains uncertain. In addition, Doppler BP measurement cannot be performed by patients at home, requiring trained personnel in the hospital setting. Our results indicate that (1) contrary to previous reports, Doppler BP better reflects systolic BP rather than mean arterial pressure; (2) Terumo (an automated BP device with a slow cuff-deflation setting to enhance sensitivity) has a high rate of measurement success, particularly in comparison to traditional automated BP devices; and (3) Terumo BP monitor provides an accurate and reliable measurement of systolic BP and of mean arterial pressure. The data from this study may allow physicians to more accurately optimize medical treatment both in the inpatient and outpatient setting. Further studies are warranted to test whether home BP monitoring may translate into fewer hypo- and hypertension-related events and, potentially, into a higher rate of myocardial recovery by enabling faster uptitration of neurohormonal blockade.