Pacemaker Optimization in Nonresponders to Cardiac Resynchronization Therapy: Left Ventricular Pacing as an Available Option

Pacemaker Optimization in Nonresponders to CardiacResynchronization Therapy: Left Ventricular Pacing asan Available OptionRYAN M. GAGE, M.S., KEVIN V. BURNS, PH.D., DANIEL B. VATTEROTT, B.A.,SPENCER H. KUBO, M.D., and ALAN J. BANK, M.D.United Heart & Vascular Clinic, St. Paul, Minnesota

Background: Echocardiographic (ECHO)-guided pacemaker optimization (PMO) in cardiac resynchro-nization therapy (CRT) nonresponders acutely improves left ventricular (LV) function. However, thechronic results of LV pacing in this group are less understood.

Methods: We retrospectively studied 28 CRT nonresponders optimized based on ECHO to LV pacingand compared them to 28 age- and gender-matched patients optimized to biventricular (BiV) pacing.ECHOs with tissue Doppler imaging assessed LV hemodynamics before, immediately after, and 29 ± 16months after PMO. Also, 56 age- and gender-matched CRT responders were included for comparison ofclinical outcomes.

Results: PMO resulted in acute improvements in longitudinal LV systolic function and several measuresof dyssynchrony, with greater improvements in the LV paced group. Chronic improvements in ejectionfraction (EF) (3.2 ± 7.7%), and left ventricle end-systolic volume (LVESV) (−11 ± 36 mL) and onedyssynchrony measure were seen in the combined group. Chronically, both LV and BiV paced patientsimproved some measures of systolic function and dyssynchrony although response varied between thegroups. Survival at 3.5 years was similar (P = 0.973) between the PMO (58%) and nonoptimized groups(58%) but survival free of cardiovascular hospitalization was significantly (P = 0.037) better in thenonoptimized group.

Conclusions: CRT nonresponders undergoing PMO to either LV or BiV pacing have acute improvementsin longitudinal systolic function and some measures of dyssynchrony. Some benefits are sustainedchronically, with improvements in EF, LVESV, and dyssynchrony. A strategy of ECHO-guided PMO resultsin survival for CRT nonresponders similar to that of CRT patients not referred for PMO. (PACE 2012;35:685–694)

heart failure, cardiac resynchronization therapy, pacemaker optimization, left ventricular pacing

IntroductionFor over a decade cardiac resynchronization

therapy (CRT) has been shown to decreasehospitalization and mortality rates in heart failure(HF) patients.1–4 Additionally, CRT improvesHF through increasing left ventricular (LV)systolic function,5,6 decreasing LV size,5–7 de-creasing intraventricular LV dyssynchrony,8 and

Address for reprints: Ryan Gage, M.S., United Heart & VascularClinic, 225 N. Smith Ave. #400, St. Paul, MN 55102. Fax: 651-241-2818; e-mail: [email protected]

Funding: This work was funded by a research grant fromMedtronic, Inc., Minneapolis, MN.

Conflict of interest: Alan Bank performs consulting andreceives research grant support from Medtronic, BostonScientific, Biotronik, and St. Jude Medical. Ryan Gage,Kevin Burns, Dan Vatterott, and Spencer Kubo and have nodisclosures.

Received August 25, 2011; revised January 17, 2012; acceptedJanuary 24, 2012.

doi: 10.1111/j.1540-8159.2012.03384.x

increasing exercise capacity.3 However, 25–35%of CRT patients meeting class 1 indications(ejection fraction [EF] ≤ 35%, QRS > 120 ms,and New York Heart Association [NYHA] classIII/IV clinical symptoms on optimal medicaltherapy) are considered nonresponders based onechocardiographic (ECHO) or clinical criteria.9,10

Pacemaker optimization (PMO) protocols havebeen developed in an attempt to convert initialnonresponders of CRT to chronic responders.11–16

Most of these studies have investigated theacute hemodynamic or ECHO effects of PMO viachanging atrioventricular (A-V) or interventricular(V-V) timing. In this study, we retrospectivelyevaluate an ECHO-guided optimization in whichLV pacing was one strategy employed to improveLV mechanics in select CRT nonresponders. Weassess the chronic ECHO and clinical responseto PMO in those patients in which LV pacingacutely improved LV function, as well as those inwhich more traditional biventricular (BiV) pacingproduced optimal LV function.

C©2012, The Authors. Journal compilation C©2012 Wiley Periodicals, Inc.

PACE, Vol. 35 June 2012 685

GAGE, ET AL.

MethodsPatient Selection

We retrospectively analyzed data from allpatients in our clinic referred for PMO duringthe years 2003–2008 who were optimized basedon ECHO findings to LV pacing (n = 28). Inaddition, we studied 28 age- and gender-matchedpatients during this same time period who wereoptimized based on ECHO findings to some formof BiV pacing. These patients were referred forPMO study by their cardiologist because of aweak or absent clinical and/or ECHO responseto CRT. In addition, we studied 56 age- andgender-matched CRT patients not referred forPMO (predominantly responders to CRT) in orderto compare survival and survival free of CVhospitalization rates in the combined optimizedgroup (LV and BiV) versus the nonoptimizedgroup. On average, these 56 CRT responders hada significant increase in EF comparing pre-CRTECHOs to studies approximately 1 year post-CRT(27 ± 8% vs 33 ± 9%, P = 0.0019). Electronicmedical records were utilized to abstract dataregarding device implantation, follow-up visits,medications, hospitalizations, and mortality.

ECHO-guided PMO Protocol

We have previously described the protocoland acute ECHO effects of this strategy in 50 con-secutive patients, nine of whom were treated withLV pacing as the best acute pacing modality.16 Inbrief, patients underwent a complete ECHO withtissue Doppler imaging (TDI) study to quantifysystolic function, diastolic function, and mechan-ical dyssynchrony at different device settings anddetermine the optimal settings. This protocol wasperformed by a cardiologist (AJB) working withan ECHO technician and a device technicianduring a single session of approximately 1 hour.Settings changed and assessed by ECHO in asequential and additive protocol were lower ratelimit, A-V delay, ventricular lead configuration,and V-V timing as per a previously describedalgorithm.16 Due to time restrictions during thePMO, only apical four-chamber, two-chamber, andlong-axis TDI images, as well as mitral valve(MV) inflow and regurgitation (MR), were seriallyassessed after altering pacemaker settings. Theoptimal pacing modality with respect to systolicfunction, diastolic function, and/or dyssynchronywas determined by the cardiologist supervisingthe study and this modality was programmed.The acute and chronic effects of PMO reportedin this paper were measured at these optimizedpacemaker settings. In the patients programmedto LV pacing, this configuration was set becausethe patient had initial negative (abnormal) lon-

gitudinal displacement of walls near the rightventricular (RV) pacing lead (inferior, septal,or anteroseptal walls) which was improved byturning off the RV lead (Fig. 1). In patients whoseoptimal device setting was BiV, the followingchanges in pacemaker settings were made: 12 A-Vdelay, six A-V delay plus V-V timing, three lowerrate, two lower rate plus V-V timing, one A-V delayplus lower rate, one A-V delay plus lower rate plusV-V timing, and three no change.

The average time post-PMO for the chronicECHO follow-up was 22 ± 18 months for the LVpaced group and 35 ± 12 months for the BiV pacedgroup (P = 0.010). A total of 13 (6 LV, 7 BiV) of the56 patients were missing a chronic ECHO study.

ECHO with TDI studies were performed usingVivid 7 ultrasound machines (General ElectricCo., Milwaukee, WI, USA) and analyzed onthe GE EchoPAC PC reading station (version7.0.0.). EF was calculated using the modifiedSimpson’s method by tracing end-diastolic andend-systolic volumes on apical four- and two-chamber views. Severity of MR was assessed usinga semiquantitative scale of 0 to 6 (i.e. 0 = none, 1 =trace, 2 = mild, 3 = mild-moderate, 4 = moderate,5 = moderate-severe, 6 = severe).

Tissue-tracking (TT) mode was used forall TDI studies to quantify the magnitude andtiming of longitudinal displacement of regionalmyocardial segments. Apical four-chamber, two-chamber, and long-axis TDI data were acquiredat baseline, serially assessed after each newdevice setting during PMO, and at chronic follow-up. TT regions of interest (10 mm × 7 mm)were placed at six basal (septal, lateral, inferior,anterior, anteroseptal, and posterior) and six mid-ventricular segments. The starting point for allTT curves (the beginning of each heartbeat) wasdefined as the time of MV closure. Cardiac eventtimings (i.e., MV closure, aortic valve closure, etc.)were obtained by identifying these events usingstandard pulsed-Doppler ECHO of transmitral andtransaortic blood flow patterns. Events markedon pulsed-Doppler images were automaticallytransferred to TDI images by analysis software.

Global systolic contraction score (GSCS) wasused as a measure of longitudinal systolic functionand defined as the average longitudinal displace-ment at end-systole (aortic valve closure) of the12 tissue segments analyzed.17,18 Delayed onset ofactivation (DOA) was one measure of mechanicaldyssynchrony used and was defined as thenumber of segments with an initial longitudinaltissue movement (at least 0.5 mm) away fromthe transducer after MV closure.16,19 Standarddeviation of the time-to-peak displacement (TT-SD12) was used to measure systolic dyssynchronyof tissue segments.16

686 June 2012 PACE, Vol. 35

PMO IN CRT NONRESPONDERS

Figure 1. Tissue-tracking curves before pacemaker optimization (PMO), immediately after PMO, and many monthsafter PMO in two patients. Arrows demonstrate abnormal delayed onset of activation of septal (Subject 1) and inferior(Subject 2) myocardial segments that is corrected acutely and chronically following PMO.

Device Safety

Electronic medical records were reviewed foreach patient, including device checks and clinicappointments with each patient’s electrophysiol-ogist. This was performed in order to determineif there were any device malfunctions, failures,or other adverse clinical events related to deviceprogramming or function.

Subjective Clinical Response

A retrospective and subjective review ofpatient records was performed to assess clinical re-sponse (as previously described) at approximately1 year post-PMO in both LV and BiV pacedpatients.20 Patients were considered to be clinicalresponders if they were categorized as signif-icantly or mildly improved and a nonrespon-der if no changes occurred or if symptomsworsened.

Statistical Analysis

Statistical analysis and graphing were per-formed using GraphPad Prism version 4.00 for

Windows (GraphPad Software, San Diego, CA,USA). Variables measured at baseline, post-PMO,and chronically were analyzed by analysis ofvariance for within-group differences. Between-group differences at the same time point wereaccessed by two-sample t-tests. Between-groupproportional differences in medications, leadlocations, and intraventricular conduction delayblocks were measured using Fischer’s Exact test orχ2-test. Between-group comparison of subjectiveclinical response was tested using a 4 × 2contingency table and χ2-test.

Kaplan-Meier curves were calculated and as-sessed with the Log-Rank test comparing survivaland survival free of CV hospitalization. The twogroups of CRT nonresponders (LV and BiV paced)were combined, forming an “optimized group”that was compared to the age- and gender-matchednonoptimized group. The nonoptimized group’szero time point for the Kaplan-Meier analyses wasset at 19 months postimplant, the average timeperiod between implant and PMO in the optimizedgroup. Statistical significance was defined asP < 0.05. Continuous data are presented as


GAGE, ET AL.

Table I.

Baseline Clinical Characteristics

Combined LV Group BiV Group P value

Subjects, n 56 28 28Male/Female, n 46 / 10 23 / 5 23 / 5 0.999Age at Implant, years 69 ± 10 70 ± 10 68 ± 10 0.567Duration of HF before implant – years 4.6 ± 4.9 4.5 ± 4.1 4.8 ± 5.8 0.861Months from Implant to PMO 19 ± 16 19 ± 16 19 ± 16 0.888Months from PMO to chronic ECHO 29 ± 16 22 ± 18 35 ± 12 0.010EF before Implant (%) 24 ± 6.5 23 ± 5.9 26 ± 6.9 0.160QRS Duration (ms) 140 ± 33 135 ± 31 146 ± 35 0.244HF Etiology, n (%)

Ischemic 39(70) 19(68) 20(71) 0.999Valvular 3(5) 3(11) 0(0) 0.236Idiopathic 14(25) 6(21) 8(29) 0.759

HF Medications n (%)ACE/ARB 44(79) 20(71) 24(86) 0.329β-Blocker 47(84) 23(82) 24(86) 0.999Aspirin 32(57) 19(68) 13(46) 0.177Digoxin 28(50) 16(57) 12(43) 0.423Anti-Arrhythmic 14(25) 6(21) 8(29) 0.759Aldosterone Blocker 11(20) 6(21) 5(18) 0.999

RV Lead Position n (%)* 0.191Septal 36(67) 19(73) 17(61)High 6 3 3Mid 27 16 11Low 3 0 3Apical 18(33) 7(27) 11(39)

LV Lead Position n (%)* 0.324Lateral 32(59) 15(58) 17(61)Posterior 2(4) 2(8) 0(0)Posterior-Lateral 20(37) 9(34) 11(39)

QRS Configuration n (%)† 0.239RBBB 1(2) 1(4) 0(0)LBBB 21(42) 9(35) 12(50)IVCD 14(28) 10(38) 4(17)Paced 14(28) 6(23) 8(33)

% V-paced at optimization 97 ± 4.8 96 ± 5.8 98 ± 3.4 0.133

Continuous values are mean ± standard deviation. Binary measures are count (within-group percentage). P-values compare LV to BiV.HF = heart failure; PMO = pacemaker optimization; EF = ejection fraction; RV = right ventricle; LV = left ventricle; RBBB = right bundlebranch block; LBBB = left bundle branch block; IVCD = nonspecific intraventricular conduction delay.*Two LV patients did not have implant records indicating where leads were placed.†Two LV and four BiV patients did not have a preimplant EKG, or mention of block in physician notes.

mean ± standard deviation. The protocol wasapproved by an Institutional Review Board.

ResultsBaseline Characteristics

Device interrogation before PMO indicated anaverage of 97 ± 4.8% BiV pacing. All patientswere at least 2 months after device placement,averaging 19 ± 16 months between implantation

and PMO. Baseline demographic, medication, anddevice implantation data for patients receivingPMO are shown in Table I. Optimized patientswere 69 ± 10 years old, had a QRS duration of140 ± 33 ms, and an EF of 24 ± 6.5% before CRT.There were no significant differences in baselinecharacteristics between the LV and BiV groups.EF before CRT implant tended to be greater inthe BiV group, while a greater proportion of LVpacing patients tended to be on aspirin, but neither



Table II.

Echocardiographic Characteristics

Combined LV Group BiV Group

HR (bpm)Baseline 68 ± 13 70 ± 13 68 ± 9Acute −3 ± 6* −2 ± 7 −3 ± 6*Chronic −1 ± 10 0 ± 10 −1 ± 10EF (%)Baseline 28.1 ± 6.2 28.0 ± 5.8 28.1 ± 6.6Acute 0.2 ± 3.4 0.7 ± 3.5 −0.3 ± 3.2Chronic 3.2 ± 7.7* 1.3 ± 6.6 5.2 ± 8.4*GSCS (mm)Baseline 3.7 ± 1.7 3.4 ± 1.5 3.9 ± 1.9Acute 0.8 ± 1.0* 1.1 ± 0.9* 0.6 ± 1.0*Chronic 0.5 ± 1.8 0.7 ± 1.5* 0.2 ± 2.1LVEDV (mL)Baseline 160 ± 45 157 ± 40 163 ± 50Acute −2 ± 14 −1 ± 15 −3 ± 14Chronic −11 ± 42 −2 ± 35 −20 ± 48LVESV (mL)Baseline 116 ± 38 114 ± 36 118 ± 41Acute −2 ± 12 −3 ± 13 −1 ± 11Chronic −11 ± 36* −3 ± 28 −20 ± 41*MR (score)Baseline 2.0 ± 1.2 2.3 ± 1.2 1.8 ± 1.2Acute 0.1 ± 8 0.1 ± 0.8 0.0 ± 0.8Chronic 0.4 ± 1.2* 0.6 ± 1.3* 0.3 ± 1.1MV VTI (cm)Baseline 19.2 ± 5.8 19.5 ± 6.7 19.0 ± 4.7Acute 0.6 ± 3.4 −0.3 ± 3.4 1.5 ± 3.2*Chronic 0.2 ± 5.5 0.9 ± 5.6 −0.5 ± 5.3DOA (#)Baseline 2.6 ± 2.5 3.4 ± 2.8 1.9 ± 2.1Acute −0.6 ± 2.2* −0.8 ± 2.6 −0.4 ± 1.8Chronic −1.2 ± 3.3* −1.6 ± 3.7 −0.8 ± 3.0TT-SD12 (ms)Baseline 80.3 ± 29.5 80.5 ± 28.0 80.1 ± 31.5Acute −9.3 ± 20.9* −16.0 ± 21.9* −2.8 ± 17.9Chronic −6.9 ± 28.3 −12.4 ± 26.3* −0.6 ± 30.0

Acute and chronic values are changes from baseline. Values aremean ± standard deviation.HR = heart rate; EF = ejection fraction; GSCS = global systoliccontraction score; LVEDV = left ventricular end-diastolic volume;LVESV = left ventricular end-systolic volume; MR = mitralregurgitation; MV VTI = mitral valve velocity time integral;DOA = delayed onset of activation; TT-SD12 = standarddeviation of time-to-peak for 12 longitudinal tissue trackingsegments.*Indicates within-group P < 0.05 compared to baseline value.

was significantly different. Baseline ECHO dataobtained at the time of PMO is shown in Table II.Patients had dilated ventricles (left ventricularend-diastolic volume = 160 ± 45 mL), depressedsystolic function (EF = 28.1 ± 6.2%), and longitu-dinal dyssynchrony (TT-SD12 = 80.3 ± 29.5 ms).

Between group comparison demonstrated onlyone significant difference at baseline, with theLV pacing group having a greater DOA whencompared to the BiV group (3.6 ± 2.8 vs 1.9 ±2.1 wall segments, P = 0.026).

Acute Echocardiographic Changes

Acute and chronic ECHO changes followingPMO are compared to baseline measurementsin Table II. Patients receiving PMO had acuteimprovements in GSCS (0.8 ± 1.0 mm, P <0.0001), DOA (–0.6 ± 2.2, P = 0.044), and TT-SD12(–9.3 ± 20.9 ms, P = 0.002). Both pacinggroups had significant within-group increasesin longitudinal systolic function as quantifiedby GSCS, with the increase in the LV pacinggroup being significantly greater than that of theBiV group (1.1 ± 0.9 mm vs 0.6 ± 1.0 mm,P = 0.025). The LV pacing group also had asignificantly greater reduction in dyssynchrony(TT-SD12) when compared to the BiV group(–16.0 ± 21.9 ms vs –2.8 ± 17.9 ms, P = 0.018).Lastly, the acute changes in MV velocity timeintegral (MV VTI) tended to be greater in the BiVgroup (1.5 ± 3.2 cm vs –0.3 ± 3.4 cm, P = 0.060).There were no acute changes in EF, LV volumes,or MR severity compared to pre-PMO. There werealso no differences in ECHO measures of responsebetween patients with septal (n = 37) and thosewith apical (n = 17) RV leads.

Chronic Echocardiographic Changes

Chronic ECHO analysis revealed that areduction in DOA (–1.2 ± 3.3, P = 0.032) wassustained for those receiving PMO. In addition,chronic improvements were found in EF (3.2 ±7.7%, P = 0.009) and LVESV (–11 ± 36 mL, P =0.046). Severity of MR was observed to increasechronically (0.4 ± 1.2, P = 0.027) in the combinedgroup. There were no chronic differences inresponse between the two patient groups, althoughthere were several significant differences withineach group compared to baseline. The BiV grouphad an increase in EF (5.2 ± 8.4%, P = 0.010) anda decrease in LVESV (–20 ± 41 mL, P = 0.035).The LV paced group showed an increase in MRseverity (0.6 ± 1.3, P = 0.050), decrease in TT-SD12 (–12.4 ± 26.3 ms, P = 0.043), and an increasein GSCS (0.7 ± 1.5 mm, P = 0.042). Changein DOA was found to be strongly correlated tochange in GSCS (r = –0.4757, P = 0.003), while nosignificant correlation (r = 0.124, P = 0.4574) wasfound between change in DOA and change in TT-SD12. No differences in chronic ECHO measuresof response were observed between patients withthe two RV pacing lead locations.


GAGE, ET AL.

Figure 2. Kaplan-Meier estimates of survival between CRT nonresponders referred for a PMOand a “nonoptimized” group of patients not referred for PMO because of benefits from CRT.Survival rates are similar between the two groups. PMO vs No PMO, P = 0.973.

Figure 3. Kaplan-Meier estimates of survival between CRT nonresponders referred for a PMOand a “nonoptimized” group of patients not referred for PMO because of benefits from CRT.The group of patients not clinically referred for a PMO experienced a greater survival free of CVhospitalization. PMO vs No PMO, P = 0.037.

Subjective Clinical Response

A clinical response was determined at roughly1 year post-PMO in 23/28 patients in each of thetwo groups. The LV-only group had two patientssignificantly improve, 11 mildly improve, sevenshow no change, and three worsen, while theBiV paced patients had nine patients significantlyimprove, six mildly improve, six show no change,and two worsen. The clinical responder rate (LV-only 57% vs BiV 65%, P = 0.102) was similarbetween the two groups.

Survival and Survival Free of CV Hospitalization

Survival in the combined (LV and BiV)PMO group was 58% at 3.5 years, which wasno different (P = 0.973) than survival in thenonoptimized control group (Fig. 2). There wasa significant difference (P = 0.037) in survivalfree of CV hospitalization between the optimizedand nonoptimized groups (44% vs 27% over3.5 years) in favor of the nonoptimized patients(Fig. 3).



Device SafetyEvents of device failure or malfunction

occurred three times in the LV paced groupand twice in the BiV group. All but one of theevents involved leads that exhibited high pacingthresholds, were damaged, or created diaphragmstimulation, with only a single event being linkedto device generator failure. There were no adverseclinical events related to device safety.

DiscussionThe main finding of this study is that

an ECHO/TDI-guided PMO strategy for CRTnonresponders, including the use of LV pacingin select patients, results in both acute andchronic improvements in some measures of LVsize, systolic function, and dyssynchrony. Thisstrategy is safe and produces survival rates inoptimized patients that are similar to those inmatched patients who were not referred for PMOstudy due to positive response to CRT.

Acute Effects of PMO

A number of reports have been publisheddescribing techniques for optimizing cardiac func-tion following CRT.21 Optimization of A-V delayhas been performed using various formulas,22,23

iterative ECHO methods based on mitral inflowpatterns,24–26 measurement of mitral or aorticVTI,24,25 finger plethysmography,27–29 impedancecardiography,26 and device-based algorithms.30

Many of these studies have shown acute improve-ment in LV systolic function, diastolic function,and/or dyssynchrony. Similarly, optimization ofV-V timing has been performed using aortic VTI,31

TDI,16 radionuclide ventriculography,32 invasivemeasurement of LV dP/dt,12 finger plethysmog-raphy,28 and device algorithms.31 In the presentstudy, we describe the acute hemodynamic effectsof an ECHO-guided PMO protocol. We havepreviously shown the results of this optimizationstrategy in 50 consecutive CRT nonresponders.16

In this paper we extend our previous findings byshowing chronic improvements in measurementsof both LV synchrony and systolic function whilecomparing the effects in patients optimized basedon ECHO findings to BiV and LV pacing.

Left Ventricular Pacing

The acute hemodynamic benefits of LVpacing have been studied and shown to becomparable to those achieved in BiV pacing innonhuman subjects33,34 and in small groups ofpatients.35,36 Three larger multicenter studies havereported conflicting results about the effectivenessof LV pacing. Decrease-HF reported that thebenefits in LV size and EF were similar be-

tween sequential BiV, simultaneous BiV, and LVpacing.37 Similarly, the BELIEVE study reportedreductions in LV size (–25.5 mL vs –19.7 mL,P = 0.617) and increases in EF (4.2% vs 5.2%,P = 0.701) to be similar when comparing BiVand LV pacing.38 Different results were reported inthe B-LEFT HF trial. This 1:1 randomized trial ofpatients with idiopathic dilated cardiomyopathyreported that both EF (12.5% vs 5.1%, P = 0.01)and LV size (8.69 mm vs 5.1 mm, P = 0.05)improvements were much greater in the BiVgroup as compared to the LV pacing group.39

A mechanistic study investigating the effect ofintrinsic conduction of the right bundle and LVpacing response was performed by van Gelderet al.40 Invasive LV dP/dt was measured duringLV and BiV pacing in 34 patients meeting standardCRT implantation criteria, with ventricular fusionconfirmed by a reduction in duration and changein morphology of the QRS complex. This studyshowed that when LV pacing achieved fusionwith intrinsic right bundle conduction an acutesuperior (P = 0.0009) hemodynamic response wasattained as compared to BiV pacing. However,when fusion was not permitted because of pacingat shorter AV delays, BiV pacing was needed toachieve a better acute hemodynamic response.In the current study we were able to measuresingle-lead EKGs acquired during PMO in 25 of28 LV paced patients. Of these, presumed fusionoccurred in 11 patients as the QRS complexshortened with a morphology change during LVpacing.

This study is very different from those mul-ticenter studies described above, as patients wereprospectively randomized to receive LV pacing atthe time of their CRT implantation. We studiedthis pacing modality only in CRT nonresponders.In addition, we individually tailored our therapyto specific patients who had TDI dyssynchronypatterns similar to those we have observed inpatients with HF and dyssynchrony from RVpacing.19 As shown in Fig. 1A, patients optimizedto LV pacing typically had an initial abnormalnegative longitudinal motion of the myocardialsegments (septal, inferior, and/or anteroseptal)near the RV pacing lead. We have called thisDOA since there is a delay in the onset of normalmyocardial displacement toward the base of theheart and away from the apex. Prinzen et al.and Peschar et al. have shown in a numberof animal studies that RV pacing produces thegreatest structural and functional abnormalities inthe myocardial segments adjacent to the RV pacinglead.41,42 These abnormalities include thinningof the wall, reduced myocardial blood flow,decreased fiber shortening, impaired contractilework and oxygen consumption, and regional wall


GAGE, ET AL.

motion abnormalities.41–43 In human subjects withnormal LV systolic function, we have shownthat acute RV apical and BiV pacing (with anapical RV lead) results in abnormal longitudinalmotion of walls near the RV pacing lead.44 Whenthe RV lead is turned off (Fig. 1B) longitudinalmotion is normalized, and this hemodynamicimprovement can persist chronically (Fig. 1C).To our knowledge, there are no previous studiesassessing LV pacing as a strategy for treating selectCRT nonresponders.

In addition to demonstrating ECHO improve-ments with LV pacing in this select group ofpatients, we show that this strategy is safe. We didnot have any adverse effects as a result of turningoff the RV lead. Since the patients undergoingPMO were already an average of 19 months post-CRT, the LV lead was very likely in stable positionand unlikely to lose capture. However, we triedto avoid LV pacing in patients with underlyingcomplete heart block since these patients aredependent upon ventricular pacing and at risk ofsyncope if LV lead failure occurred.

Chronic Effects of PMO

There are few studies of PMO that havereported chronic outcomes. Taha et al. foundthat comprehensive echo-guided PMO resulted inacute improvements in Doppler hemodynamicsand S3 intensity.11 NYHA class and the physicalcomponent of the Minnesota Living with HeartFailure Questionnaire score improved at 1 monthpostoptimization. These findings were limitedby the short follow-up period, significant lossof patients to follow-up, and the lack of acontrol group. The In-Sync III study was amulticenter nonrandomized prospective studyevaluating sequential BiV pacing in 422 patientswith moderate-to-severe HF.45 A-V delay wasprogrammed based on the Ritter method and V-V delay was programmed based on cardiac outputdetermined from ECHO Doppler measurement ofLV outflow tract velocities. Six-month follow-up demonstrated modest improvements in strokevolume and 6-minute hall walk distance but noimprovements in quality of life or functionalstatus.

Our study provides unique information on thechronic effects of PMO. In our combined groupof 56 optimized patients, EF and LVESV improvesignificantly over a mean follow-up period ofalmost 2.5 years. These findings are particularlyimportant in that they occurred in patients whowere referred for PMO due to nonresponse toCRT. One would expect this group to be verydifficult to treat and successfully sustain LVreverse remodeling. After PMO, this group ofnonresponders to CRT demonstrated a mortality

rate (approximately 12% per year) that is similarto matched patients who received CRT but werenot referred for PMO because of clinical responseto CRT. This mortality rate is slightly higher thanthe approximately 10.6% mortality found in theMedicare Implantable Cardioverter-DefibrillatorRegistry of 14,946 patients.46 However, the registryincluded 12% NYHA Class I and II patients whoare well known to have much lower mortalitythan the Class III and IV patients in our study.Survival free of cardiovascular hospitalizationwas significantly worse in our PMO patientsas compared to the control nonoptimized pa-tients. This is not surprising since the patientsundergoing PMO were preselected as havinga poor prognosis post-CRT due to referral asnonresponders.

Limitations

This study was retrospective, nonrandom-ized, and contained data from only a singlecenter. We do not have a control group thatwas referred for PMO but did not have theirpacemaker settings optimized. The definitionof nonresponse to CRT varies across studies.In our retrospective study, CRT nonresponsewas subjectively determined by the referringcardiologist. Since the mean time to PMO was 19months, the patients had a long time period ofobservation and drug titration before being labeledas nonresponders. This study had a relativelysmall sample size, making comparisons of survivaland CV hospitalizations less conclusive. However,as previously mentioned, there are little chronicdata on these outcomes in CRT nonresponderstreated with PMO. In addition, the mean lengthof follow-up was 4 ± 2 years, which is longer thanmost large multicenter CRT trials.1–4

ConclusionsLV pacing in select CRT nonresponders re-

sults in acute improvements in LV synchrony andsystolic function that are greater than those foundin CRT nonresponders optimized to BiV pacing.LV pacing also results in chronic improvementsin some measures of LV synchrony and systolicfunction, but optimized BiV patients trendedtoward greater improvements in LV volumesand EF chronically. Survival, but not survivalfree of cardiovascular hospitalization, followingPMO is similar in optimized nonresponders ascompared with patients not referred for PMOdue to initial response to CRT. We conclude thatECHO-guided PMO with LV or BiV pacing inselect patients is a safe and reasonable strategy fortreating CRT nonresponders that warrants furtherinvestigation.



References1. Bristow MR, Saxon LA, Boehmer J, Kruger S, Kass DA, De Marco T,

Carson P, et al. Cardiac-resynchronization therapy with or withoutan implantable defibrillator in advanced chronic heart failure. NEng J of Med 2004;350:2140–2150.

2. Cleland JG, Daubert JC, Erdmann E, Freemantle N, Gras D,Kappenberger L, Tavazzi L. The effect of cardiac resynchronizationon morbidity and mortality in heart failure. N Eng J of Med2005;352:1539–1549.

3. Abraham WT, Fisher WG, Smith AL, Delurgio DB, Leon AR, LohE, Kocovic DZ, et al. Cardiac resynchronization in chronic heartfailure. N Engl J of Med 2002;346:1845–1853.

4. Cazeau S, Leclercq C, Lavergne T, Walker S, Varma C, Linde C,Garrigue S, et al. Effects of multisite biventricular pacing in patientswith heart failure and intraventricular conduction delay. N Eng Jof Med 2001;344:873–880.

5. St. John Sutton MG, Plappert T, Abraham WT, Smith AL, DeLurgioDB, Leon AR, Loh E, et al. Effect of cardiac resynchronizationtherapy on left ventricular size and function in chronic heartfailure. Circulation 2003;107:1985–1990.

6. St. John Sutton MG, Plappert T, Hilpisch KE, Abraham WT,Hayes DL, Chinchoy E. Sustained reverse left ventricular structuralremodeling with cardiac resynchronization at one year is a functionof etiology. Circulation 2006;113:266–272.

7. Solomon SD, Foster E, Bourgoun M, Shah A, Viloria E, BrownMW, Hall J, et al. Effect of cardiac resynchronization therapy onreverse remodling and relation to outcome multicenter automaticdefibrillator implantation trial: Cardiac resynchronization therapy.Circulation 2010;122:985–992.

8. Yu CM, Chau E, Sanderson JE, Katherine F, Man-Oi T, FungWH, Lin H, et al. Tissue Doppler echocardiographic evidence ofreverse remodeling and improved synchronicity by simultaneouslydelaying regional contraction after biventricular pacing therapy inheart failure. Circulation 2002;105438–445.

9. Birnie DH, Tang ASL. The problem of non-response to cardiacresynchronization therapy. Curr Opin in Cardiol 2006;21:20–26.

10. Yu CM, Fung JWH, Zhang Q, Sanderson JE. Understanding nonre-sponders of cardiac resynchronization therapy—Current and futureperspectives. J Cardiovasc Electrophysiol 2005;16:1117–1124.

11. Taha N, Zhang J, Ranjan R, Daneshvar S, Castillo E, Guillen E,Montoya MC, et al. Biventricular pacemaker optimization guidedby comprehensive echocardiography—Preliminary observationsregarding the effects on systolic and diastolic ventricularfunction and third heart sound. J Am Soc Echocardiography2010;23:857–866.

12. van Gelder BM, Meijer A, Bracke FA. The optimized V-V intervaldetermined by interventricular conduction times versus invasivemeasurement by LV dP/dt MAX. J Cardiovasc Electrophysiol2008;19:939–944.

13. Naqvi TZ, Rafique AM. Echocardiography-guided pacemakeroptimization and radial artery tonometry. J Card Fail 2008;14:583–589.

14. Naqvi TZ. Echocardiography-guided biventricular pacemakeroptimization. J Am Coll Cardiol: Cardiovasc Imag 2010;3:1168–1180.

15. Mullens W, Grimm RA, Verga T, Dresing T, Starling RC, WilkoffBL, Tang WHW. Insights from a cardiac resynchronizationoptimization clinic as part of a heart failure disease managementProgram. J Am Coll Cardiol 2009;53:765–773.

16. Bank AJ, Burns KV, Kelly AS, Thelen AM, Kaufman CL, Adler SW.Echocardiographic improvements with pacemaker optimization inthe chronic post cardiac resynchronization therapy setting. ClinMed: Cardiol 2008;2:65–74.

17. Søgaard P, Engeblad H, Pedersen AK, Kim WY, KristensenBØ, Hansen PS, Mortensen PT. Sequential versus simultaneousbiventricular resynchronization for severe heart failure. Circulation2002;106:2078–2084.

18. Søgaard P, Engeblad H, Kim WY, Jensen HK, Pedersen AK,Kristensen BØ, Mortensen PT. Tissue Doppler imaging predictsimproved systolic performance and reversed left ventricularremodeling during long-term cardiac resynchronization therapy.J Am Coll Cardiol 2002;40:723–730.

19. Bank AJ, Kaufman CL, Burns KV, Parah JS, Johnson L, KellyAS, Shroff SG, et al. Intramural dyssynchrony and response tocardiac resynchronization therapy in patients with and withoutprevious right ventricular pacing. Eur J of Heart Fail 2010;12:1317–1324.

20. Bank AJ, Burns KV, Gage RM, Vatterott DB, Adler SW, SajadyM, Rohde D, et al. Cardiac resynchronization therapy in the realworld: Comparison with the COMPANION study. J Card Fail2012;18:153–158.

21. Burri H, Sunthorn H, Shah D, Lerch R. Optimization of deviceprogramming for cardiac resynchronization therapy. Pacing ClinElectrophysiol 2006;29:1416–1425.

22. Ritter P, Dib JC, Lelievre T. Quick determination of the optimal AVdelay at rest in patients paced in DDD mode for complete AV block.Eur J Card Pacing Electrophysiol 1994;4:39.

23. Ishikawa T, Sumita S, Kimura K, Kikuchi M, Kosuge M,Kuji N, Endo T, et al. Prediction of optimal atrioventriculardelay in patients with implanted DDD pacemakers. Pacing ClinElectrophysiol 1999;22:1365–1371.

24. Gold MR, Niazi I, Giudici M, Leman RB, Sturdivant JL, KimMH, Yu Y, et al. A prospective comparison of AV delayprogramming methods for hemodynamic optimization duringcardiac resynchronization therapy. J Cardiovasc Electrophysiol2007;18:490–496.

25. Jansen AHM, Bracke FA, van Dantzig JM, Meijer A, van der VoortPH, Aarnoudse W, van Gelder BM, et al. Correlation of Echo-Doppler optimization of atrioventricular delay in cardiac resyn-chronization therapy with invasive hemodynamics in patientswith heart failure secondary to ischemic or idiopathic dilatedcardiomyopathy. Am J Cardiol 2006;97:552–557.

26. Kindermann M, Frohlig G, Doerr T, Schieffer H. Optimizing theAV delay in DDD pacemaker patients with high degree AV block:Mitral valve Doppler versus impedance cardiography. Pacing ClinElectrophysiol 1997;20:2453–2462.

27. Whinnett ZI, Davies JER, Willson K, Chow AW, Foale RA, DaviesDW, Hughes AD, et al. Determination of optimal atrioventriculardelay for cardiac resynchronization therapy using acute non-invasive blood pressure. Europace 2006;8:358–366.

28. Whinnett ZI, Davies JER, Willson K, Manisty CH, ChowAW, Foale RA, Davies DW, et al. Haemodynamic effects ofchanges in atrioventricular and interventricular delay in cardiacresynchronization therapy show a consistent pattern: Analysis ofshape, magnitude and relative importance of atrioventricular andinterventricular delay. Heart 2006;92:1628–1634.

29. Butter C, Stellbrink C, Belalcazar A, Villalta D, Schlegl M, SinhaA, Cuesta F, et al. Cardiac resynchronization therapy optimizationby finger plethysmography. Heart Rhythm 2004;1:568–575.

30. Ellenbogen KA, Gold MR, Meyer TE, Lozano IF, Mittal S,Waggoner AD, Lemke B, et al. Primary results from theSmartDelay determined AV optimization: A comparison to otherAV delay methods used in cardiac resynchronization therapy(SMART-AV) trial: A randomized trial comparing empirical,echocardiography-guided, and algorithmic atrioventricular delayprogramming in cardiac resynchronization therapy. Circulation2010;122:2660–2668.

31. Baker JH, McKenzie J, Beau S, Greer GS, Porterfield J, FedorM, Greenberg S, et al. Acute evaluation of programmer-guidedAV/PV and VV delay optimization comparing an IEGM methodand echocardiogram for cardiac resynchronization therapy in heartfailure patients and dual-chamber ICD implants. J CardiovascElectrophysiol 2007;18:185–191.

32. Burri H, Sunthorn H, Somsen A, Zaza S, Fleury E, Shah D, RighettiA. Optimizing sequential biventricular pacing using radionuclideventriculography. Heart Rhythm 2005;2:960–965.

33. Helm RH, Leclercq C, Faris OP, Ozturk C, McVeigh E, Lardo AC,Kass DA. Cardiac dyssynchrony analysis using circumferentialversus longitudinal strain: Implications for assessing cardiacresynchronization. Circulation 2005;111:2760–2767.

34. Leclercq C, Faris O, Tunin R, Johnson J, Kato R, Evans F, SpinelliJ, et al. Systolic improvement and mechanical resynchronizationdoes not require electrical synchrony in the dilated failing heartwith left bundle-branch block. Circulation 2002;106:1760–1763.

35. Nelson GS, Berger RD, Fetics BJ, Talbot M, Spinelli JC, HareJM, Kass DA. Left ventricular or biventricular pacing improvescardiac function at diminished energy cost in patients withdilated cardiomyopathy and left bundle-branch block. Circulation2000;102:3053–3059.

36. Kawaguchi M, Murabayashi T, Fetics BJ, Nelson GS, SamejimaH, Nevo E, Kass DA. Quantitation of basal dyssynchrony andacute resynchronization from left or biventricular pacing by novelecho-contrast variability imaging. J Am Coll Cardiol 2002;39:2052–2058.


GAGE, ET AL.

37. Rao RK, Kumar UN, Schafer J, Viloria E, De Lurgio D,Foster E. Reduced ventricular volumes and improved systolicfunction with cardiac resynchronization therapy: A randomizedtrial comparing simultaneous biventricular pacing, sequentialbiventricular pacing, and left ventricular pacing. Circulation2007;115:2136–2144.

38. Gasparini M, Bocchiardo M, Lunati M, Ravazzi PA, Santini M,Zardini M, Signorelli S, et al. Comparison of 1-year effects ofleft ventricular and biventricular pacing in patients with heartfailure who have ventricular arrhythmias and left bundle-branchblock: The Bi vs left ventricular pacing: An international pilotevaluation on heart failure patients with ventricular arrhythmias(BELIEVE) multicenter prospective randomized pilot study. AmHeart J 2006;152:155e1–155e7.

39. Sedlacek, K, Burianova, L, Mlcochova, H, Piechl P, Marek T,Kautzner J. Isolated left ventricular pacing sesults in worse long-term clinical outcome when compared with biventricular pacing:A single-centre randomized study. Europace 2010;12:1762–1768.

40. van Gelder BM, Bracke FA, Meijer A, Pijls NHJ. The hemo-dynamic effect of intrinsic conduction during left ventricularpacing as compared to biventricular pacing. J Am Coll Cardiol2005;46:2305–2310.

41. Prinzen FW, Peschar M. Relation between the pacing inducedsequence of activation and left ventricular pump function inanimals. Pacing Clin Electrophysiol 2002;25:484–498.

42. Peschar M, de Swart H, Michels KJ, Reneman RS, Prinzen FW. Leftventricular septal and apex pacing for optimal pump function incanine hearts. J Am Coll Cardiol 2003;41:1218–1226.

43. Prinzen FW, Hunter WC, Wyman BT, McVeigh ER. Mapping ofregional myocardial strain and work during ventricular pacing:Experimental study using magnetic resonance imaging tagging. JAm Coll Cardiol 1999;33:1735–1742.

44. Bank AJ, Schwartzman DS, Burns KV, Kaufman CL, Adler SW,Kelly AS, Johnson L, et al. Intramural dyssynchrony from acuteright ventricular apical pacing in human subjects with normalleft ventricular function. J Cardiovasc Trans Research 2010;3:321–329.

45. Leon AR, Abraham WT, Brozena S, Daubert JP, Fisher WG, GurleyJC, Liang CS, et al. Cardiac resynchronization with sequentialbiventricular pacing for the treatment of moderate-to-severe heartfailure. J Am Coll Cardiol 2005;46:2298–2304.

46. Bilchick KC, Kamath S, DiMarco JP, Stukenborg GJ. Bundle-brachblock morphology and other predictors of outcome after cardiacresynchronization therapy in medicare patients. Circulation2010;122:2022–2030.


Documents

Pacemaker Optimization in Nonresponders to Cardiac Resynchronization Therapy: Left Ventricular Pacing as an Available Option