Coronary Hemodynamics
`
`Hemodynamic Response to Intravenous Adenosine and
`Its Effect on Fractional Flow Reserve Assessment
`Results of the Adenosine for the Functional Evaluation of Coronary
`Stenosis Severity (AFFECTS) Study
`
`Jason M. Tarkin, MBBS; Sukhjinder Nijjer, MBChB; Sayan Sen, MBBS; Ricardo Petraco, MD;
`Mauro Echavarria- Pinto, MD; Kaleab N. Asress, BM BCh, MA;
`Tim Lockie, MBChB, PhD; Muhammed Z. Khawaja, MBBS; Jamil Mayet, MBChB, MD;
`Alun D. Hughes, MBBS, PhD; Iqbal S. Malik, MBBS, PhD;
`Ghada W. Mikhail, MBBS, PhD; Christopher S. Baker, MBBS, PhD; Rodney A. Foale, MD;
`Simon Redwood, MBBS, MD; Darrel P. Francis, MB BChir, MA, MD;
`Javier Escaned, MD, PhD; Justin E. Davies, MBBS, PhD
`
`Background—We studied the hemodynamic response to intravenous adenosine on calculation of fractional flow reserve
`(FFR). Intravenous adenosine is widely used to achieve conditions of stable hyperemia for measurement of FFR. However,
`intravenous adenosine affects both systemic and coronary vascular beds differentially.
`Methods and Results—A total of 283 patients (310 coronary stenoses) underwent coronary angiography with FFR using
`intravenous adenosine 140 mcg/kg per minute via a central femoral vein. Offline analysis was performed to calculate
`aortic (Pa), distal intracoronary (Pd), and reservoir (Pr) pressure at baseline, peak, and stable hyperemia. Seven different
`hemodynamic patterns were observed according to Pa and Pd change at peak and stable hyperemia. The average time
`from baseline to stable hyperemia was 68.2±38.5 seconds, when both ΔPa and ΔPd were decreased (ΔPa, −10.2±10.5
`mm Hg; ΔPd, −18.2±10.8 mm Hg; P<0.001 for both). The fall in Pa closely correlated with the reduction in peripheral
`Pr (ΔPr, −12.9±15.7 mm Hg; P<0.001; r=0.9; P<0.001). ΔPa and ΔPd were closely related under conditions of peak
`(r=0.75; P<0.001) and stable hyperemia (r=0.83; P<0.001). On average, 56% (10.2 mm Hg) of the reduction in Pd was
`because of fall in Pa. FFR lesion classification changed in 9% using an FFR threshold of ≤0.80 and 5.2% with FFR
`threshold <0.75 when comparing Pd/Pa at peak and stable hyperemia.
`Conclusions—Intravenous adenosine results in variable changes in systemic blood pressure, which can lead to alterations in
`FFR lesion classification. Attention is required to ensure FFR is measured under conditions of stable hyperemia, although
`the FFR value at this point may be numerically higher. (Circ Cardiovasc Interv. 2013;6:654-661.)
`
`Key Words: adenosine ◼ angiography ◼ blood pressure ◼ coronary disease ◼ fractional flow reserve, myocardial
`◼ hemodynamics
`
`Fractional flow reserve (FFR) is widely used in clinical
`
`practice to assess severity of functional coronary artery
`stenosis; it is the ratio of distal intracoronary pressure (Pd)
`to aortic pressure (Pa) measured during pharmacologically
`induced hyperemia. FFR has been shown to be effective
`at reducing the rate of stent implantation and improving
`cardiac outcomes compared with angiographic guidance
`alone, and its use is supported by international guidelines.1,2
`Large clinical studies such as Fractional Flow Reserve ver-
`sus Angiography for Multivessel Evaluation (FAME) used
`
`Editorial see p 602
`conditions of maximal hyperemia achieved by central-line
`administration of intravenous adenosine to estimate FFR.3–5
`Other landmark studies such as DEFER used both intrave-
`nous and intracoronary adenosine, although the frequency
`of the mode of adenosine administration was not reported.3
`Intravenous adenosine administration is thought to provide
`the most stable conditions of hyperemia for measurement
`of FFR. However, even in the best hands, it is possible that
`
`Received May 7, 2013; accepted October 9, 2013.
`From the International Centre for Circulatory Health, National Heart and Lung Institute, NIHR Imperial College Biomedical Research Centre, and
`Imperial College NHS Trust, London, United Kingdom (J.M.T., S.N., S.S., R.P., J.M., A.D.H., I.S.M., G.W.M., C.S.B., R.A.F., D.P.F., J.E.D.); King’s
`College London BHF Centre of Research Excellence and the NIHR Biomedical Research Centre at Guy’s and St Thomas’ Hospital NHS Foundation
`Trust, the Rayne Institute, St Thomas’ Hospital, London, United Kingdom (K.N.A., T.L., M.Z.K., S.R.); and Cardiovascular Institute, Hospital Clinico San
`Carlos, Madrid, Spain (M.E.-P., J.E.).
`Correspondence to Justin E. Davies, MBBS, PhD, International Centre for Circulatory Health, Imperial College, 59–61 N Wharf Rd, Paddington,
`London W2 1LA, United Kingdom. E-mail justin.davies@imperial.ac.uk
`© 2013 American Heart Association, Inc.
`Circ Cardiovasc Interv is available at http://circinterventions.ahajournals.org
`
`DOI: 10.1161/CIRCINTERVENTIONS.113.000591
`
`654
`
`Downloaded from
`
`http://circinterventions.ahajournals.org/
`
` by guest on October 4, 2017
`
`CATHWORKS EXHIBIT 1012
`Page 654 of 662
`
`

`

`Tarkin et al
`
` Hemodynamic Response to Intravenous Adenosine
`
` 655
`
`WHAT IS KNOWN
`• Fractional flow reserve (FFR) has been shown to be
`effective at reducing the rate of stent implantation
`and improving cardiac outcomes compared with an-
`giographic guidance alone.
`• Clinical trial data are almost exclusively based on FFR
`values measured during maximal steady-state hyper-
`emia induced by intravenous adenosine infusion.
`
`WHAT THE STUDY ADDS
`• Intravenous adenosine results in variable and unpre-
`dictable changes in systemic blood pressure.
`• Measuring FFR prior to attainment of maximal
`steady-state hyperemia can lead to changes in clini-
`cal decision making.
`• FFR during maximal steady-state hyperemia may
`be numerically higher than at other times during
`measurement.
`• Using intravenous adenosine, on average it takes ≥60
`seconds to achieve stable pressure conditions suit-
`able for measurement of FFR.
`• The heterogeneous hyperemic response of adenosine
`compounds suggests that new agents with vastly dif-
`ferent pharmacological profiles (such as regadeno-
`son) will need further evaluation in clinical outcome
`trials prior to adoption into routine clinical practice.
`
`important changes in categorization can inadvertently occur
`when FFR values lie close to the treatment threshold (ie,
`≤0.8 or <0.75) because of small changes in the response
`to adenosine administration, possibly related to variations
`in peripheral and central hemodynamic responses. Such
`changes can frequently be observed when the initial Pd/Pa
`ratio is lower during the first moments of adenosine admin-
`istration than during stable hyperemia (Figure 1). Not only
`may initial FFR values be lower than those attained during
`stable hyperemia, but they also tend to be more variable, and
`it is difficult to interpret their meaning in the context of rig-
`orously performed FFR clinical outcome studies, which use
`the FFR values taken at maximal stable hyperemia as the
`reference standard.
`FFR should be measured under steady-state hyperemia.
`However, in clinical practice, the time at which stable hyper-
`emia is achieved can sometimes be difficult to determine, and
`often the lowest rather than stable FFR value is used for clini-
`cal decision making. This is especially apparent when relying
`on automated FFR consoles, which seek the lowest recorded
`Pd/Pa ratio rather than seeking the ratio at stable hyperemia.
`With this in mind, our study aimed to evaluate the hemody-
`namic response to intravenous adenosine. Specifically, we
`aimed to evaluate the frequency by which changes in Pd/Pa
`lead to changes in clinical categorization when FFR measure-
`ments are obtained during peak and stable hyperemia and the
`extent to which central hemodynamic changes influence coro-
`nary Pa and Pd, and the Pd/Pa.
`
`Methods
`
`Study Design
`This nonrandomized retrospective analysis included coronary catheter
`data from 283 unselected consecutive patients (310 coronary stenoses)
`who underwent coronary angiography with FFR measured using intra-
`venous adenosine as part of clinical management at 3 sites (Imperial
`College Healthcare, London, United Kingdom; Cardiovascular
`Institute, Hospital Clinico San Carlos, Madrid, Spain; St. Thomas’
`Hospital, London, United Kingdom). Cases where intracoronary ad-
`enosine was used to measure FFR were not included in the study.
`Patient demographics are reported in Table 1. All subjects gave writ-
`ten informed consent in accordance with the protocol approved by the
`local ethics committee prior to undergoing coronary angiography.
`
`Procedure and Data Acquisition
`Coronary catheter data were obtained in the standard way,6 us-
`ing a 0.014-inch pressure sensor–tipped wire (PrimeWire, Volcano
`Corporation, or Radi PressureWire, St Jude Medical, Minneapolis,
`MN) passed into the target vessel via a guiding catheter during diag-
`nostic angiography. Pressure recordings were made at baseline and
`throughout intravenous infusion of adenosine at a rate of 140 µg/kg
`per minute administered via a femoral venous sheath, and hemody-
`namic data were stored on the Combomap and Radi analyzer.
`
`Reservoir Pressure: A Measure of Systemic
`Vascular Resistance
`Reservoir pressure (Pr) is the pressure generated during systole when
`more blood is ejected from the heart than can move out to the peripheral
`vessel. During this phase, the aorta distends under pressure (Pr) to ac-
`commodate the additional systolic blood volume. Later during diastole, Pr
`declines as aortic distension recoils providing perfusion to the vasculature.
`In this way, the reservoir integrates the on–off ejection of the heart into a
`continuous blood flow throughout the entire cardiac cycle, and as such, Pr
`is a composite of compliance of the large elastic arteries (predominately
`the proximal aorta) and a measure of peripheral vascular resistance (or
`outflow to blood). Pr was calculated using the standard reservoir-wave
`separation as described previously.7 By measuring how Pr changes, it is
`possible to assess the extent to which vasodilators, such as adenosine, act
`to reduce coronary perfusion from the aortic via peripheral vasodilatation.
`
`Analysis of Hemodynamic Data
`Data were analyzed using customized software written in Matlab
`(Mathworks, Inc, Natick, MA). Baseline and stable hyperemia time
`points were confirmed by visual inspection of each hemodynamic
`trace, and peak hyperemia was defined as the lowest FFR within 60
`seconds of intravenous adenosine infusion and stable hyperemia as
`when the hyperemic plateau was reached. This analysis yielded mean
`Pa, Pd, and maximum Pr measurements at baseline, peak, and stable
`hyperemia after intravenous adenosine infusion.
`The lowest FFR during the first 60 seconds (peak FFR) was compared
`with FFR measured after confirmation of stable FFR. This was designed
`to reflect a common clinical scenario, often observed when measuring
`FFR. Automated FFR consoles are designed to track the lowest FFR val-
`ue. However, it is important that the clinician is able to visually interpret
`the hemodynamic traces and decide the time when stable hyperemia has
`been achieved to prevent the potential misinterpretation of an artifactual-
`ly low FFR reading. To avoid this potential pitfall, it is common practice
`to wait for ≥1 minute after intravenous adenosine before measuring FFR.
`
`Pressure Response and FFR Group Analysis
`Data were analyzed to determine mean pressure changes for the entire
`group in response to intravenous adenosine. Mean time from baseline
`to stable hyperemia was calculated for the entire group. Mean ∆ pres-
`sure gradient (Pa−Pd) was also compared. Stenoses were categorized
`according to direction of mean Pa and Pd change at peak and stable
`hyperemia, with reference to the previous state. Mean ∆Pa, ∆Pd, and
`∆Pr from baseline (expressed in both percentage and mm Hg) were
`
`Downloaded from
`
`http://circinterventions.ahajournals.org/
`
` by guest on October 4, 2017
`
`CATHWORKS EXHIBIT 1012
`Page 655 of 662
`
`

`

`Figure 1. Change in fractional flow reserve (FFR)
`classification between peak and stable hyper-
`emia. Aortic pressure (Pa) and distal intracoro-
`nary pressure (Pd) recordings (top) and FFR trace
`(bottom) shown during intravenous adenosine
`infusion at a rate of 140 mcg/kg per minute. Ini-
`tially, a marked fall in FFR was seen, with FFR
`moving across the 0.8 treatment threshold and
`a peak FFR being recorded of FFR=0.75. After
`reaching the nadir, the FFR value recrossed the
`0.80 treatment threshold and became stable at
`FFR=0.84.
`
`656
`
` Circ Cardiovasc Interv
`
` December 2013
`
`Pa pressure
`Pd pressure
`
`baseline
`FFR=0.92
`
`stable
`FFR=0.84
`
`0
`
`20
`
`peak
`FFR=0.75
`40
`60
`80
`Time (seconds)
`
`FFR 0.80 treatment threshold
`
`100
`
`120
`
`140
`120
`100
`80
`60
`40
`20
`0
`
`Pressure (mmHg)
`
`0.9
`
`0.85
`
`FFR
`
`0.8
`
`compared between the groups. FFR was calculated using pressure
`measurements at baseline, peak, and stable hyperemia for all stenoses.
`Changes in lesion classification between peak and stable hyperemia
`were analyzed using FFR treatment thresholds of both ≤0.80 and <0.75.
`
`Statistical Analysis
`Statistical analysis was performed using STATA version 12 (StataCorp
`LP, College Station, TX). Two-sample t test was used to compare
`pressure changes at peak and stable hyperemia with those at baseline.
`The relationship between ΔPa and ΔPd at peak and stable hyperemia
`was assessed using single-variant regression analysis and F test. To
`assess clinical characteristics in relation to likelihood of observing a
`
`Table 1. Summary of Clinical Characteristics for Total
`Study Group
`
`Patients
`Coronary stenoses
`Age, y
`Sex, % male
`Mean resting blood pressure, mm Hg
`Mean resting heart rate, beats/min
`% Luminal stenosis
`Hypertension
`Hyperlipidemia
`Diabetes mellitus
`Smoking
`Chronic renal failure
`Previous myocardial infarction
`Troponin positive
`Severe heart failure (ejection fraction <30%)
`Multivessel disease
`Previous CABG
`
`CABG indicates coronary artery bypass grafting.
`
`No.
`
`283
`310
`
`242
`
`179
`188
`104
`133
`22
`88
`24
`4
`176
`10
`
`%
`
`62.8±10.3
`85.5
`101.1±23.4
`73.6±12.2
`49±14.6
`63.2
`66.4
`36.8
`42.7
`7.8
`31.1
`8.5
`1.4
`62.2
`3.5
`
`clinical categorization change in FFR between peak and stable hyper-
`emia, first a univariant logistic regression analysis was performed and
`then variables were put into a multivariant analysis model. Logistic
`regression analysis was also used to determine the likelihood of ob-
`serving a clinical categorization change based on pressure response
`group. For all tests, P<0.05 was considered significant.
`
`Results
`
`Pressure Recordings
`Intravenous adenosine caused a significant decrease in both
`ΔPa 10.2±10.5 mm Hg (−10.0±10.1%) and ΔPd −18.2±10.8
`mm Hg (−19.8±10.3%), P<0.001 for both (Figure 2). The
`average time from baseline to stable hyperemia after adminis-
`tration of intravenous adenosine was 68.2±38.5 seconds. ΔPa
`and ΔPd were closely related under conditions of peak (r=0.75;
`P<0.001) and stable (r=0.83; P<0.001; Figure 3) hyperemia.
`On average, 56% (10.2 mm Hg) of the reduction in Pd pressure
`in response to intravenous adenosine was because of a fall in
`Pa, and the larger the fall in Pd, the larger the contribution from
`Pa (Figure 3). For example, when Pd fell by 29 mm Hg, the fall
`in Pa (23 mm Hg) accounted for 80% of the entire fall in Pd.
`At stable hyperemia, ΔPa ranged from −58.3 mm Hg
`decreased below baseline to 24.3 mm Hg increased above base-
`line. In 19.7% (61/310), Pa was decreased >20 mm Hg at stable
`hyperemia. In contrast, in 14.2% (44/310), Pa was increased
`above baseline values. The fall in Pa pressure closely followed
`the reduction in peripheral Pr, with ΔPr at stable hyperemia
`being −12.4±15.7 mm Hg (−10.4±11.9%), P<0.001. ΔPa and
`ΔPr were closely related under conditions of peak (r=0.9;
`P<0.001) and stable (r=0.9; P<0.001; Figure 4) hyperemia.
`Seven different hemodynamic patterns were observed when
`stenoses were grouped according to direction of Pa and Pd
`changes at peak and stable hyperemia (Figure 2). In 62.9%
`(17/27), hemodynamic patterns changed between different
`stenoses, within the same patient.
`
`Downloaded from
`
`http://circinterventions.ahajournals.org/
`
` by guest on October 4, 2017
`
`CATHWORKS EXHIBIT 1012
`Page 656 of 662
`
`

`

`Tarkin et al
`
` Hemodynamic Response to Intravenous Adenosine
`
` 657
`
`Figure 2. Changes in aortic pressure (Pa) and distal intracoronary pressure (Pd) at peak and stable hyperemia. Coronary stenoses
`grouped according to direction of Pa and Pd change at peak and stable hyperemia. This shows 7 different patterns of response. Mean
`∆Pa and ∆Pd (mm Hg) at baseline, peak, and stable hyperemia are shown for each group and for the total study group.
`
`FFR Measurements
`Mean resting Pd/Pa ratio (baseline) for the entire study group
`was 0.92. With infusion of intravenous adenosine, FFR fell to
`a nadir of 0.79 (peak), before rising to 0.83 with continued ade-
`nosine infusion to achieve stable hyperemia. Overall, the fall in
`FFR at peak hyperemia reflected a disproportionate fall in Pd,
`
`driven mainly by an initial relatively smaller reduction in Pa,
`leading to an increased pressure gradient and lower Pd/Pa ratio.
`At stable hyperemia, when ∆Pd more reliably reflects changes in
`coronary microcirculation, reduction in Pd was less pronounced
`relative to ∆Pa, leading to a decrease in pressure gradient and
`an increase in the Pd/Pa ratio. A similar pattern was observed
`in each of the 7 different hemodynamic groups, with FFR being
`
`Downloaded from
`
`http://circinterventions.ahajournals.org/
`
` by guest on October 4, 2017
`
`Figure 3. Relationship between change in distal intracoronary pressure (Pd) and aortic pressure (Pa) at peak and stable hyperemia. After
`administration of adenosine, both Pa and Pd fall. A significant linear relationship between fall in Pd (y axis) and Pa (x axis) after intrave-
`nous adenosine infusion was found between both peak and stable hyperemia. The reduction in Pd is attributable in part to a reduction in
`Pa and in part to vasodilatation of the coronary microcirculation.
`
`CATHWORKS EXHIBIT 1012
`Page 657 of 662
`
`

`

`658
`
` Circ Cardiovasc Interv
`
` December 2013
`
`Figure 4. Relationship between aortic pressure (Pa), distal intracoronary pressure, and reservoir pressure (Pr), at baseline, peak, and
`stable hyperemic conditions. A close relationship was found between Pa and aortic Pr at peak and stable hyperemia after intravenous
`adenosine (top). The pattern of reservoir pressure change, as a measure of peripheral resistance and large artery compliance, mirrored
`aortic pressure in each of the 7 groups (bottom).
`
`lower at peak than during stable hyperemia. However, the degree
`of change varied significantly between groups (Figure 5).
`
`Lesion Classification
`The proportion of cases where the classification changed between
`peak FFR and stable FFR was calculated for both FFR ≤0.80
`and FFR <0.75. Pd/Pa ratio was >0.8 for all stenoses at baseline.
`Using the FFR ≤0.80 cutoff, lesion classification changed in 9%
`(28/310) of cases between measurements made at peak and dur-
`ing stable FFR (Figure 1). When the FFR cutoff of <0.75 was
`used, the proportion of patients in whom classification changed
`between measurements made at peak and during stable FFR was
`5.2% (16/310). For 0.3% (1/310), FFR crossed the grey zone
`and was <0.75 at peak and >0.80 at stable hyperemia. Results of
`regression analysis comparing clinical characteristics in relation
`to likelihood of classification change in FFR between peak and
`stable hyperemia are summarized in Table 2. There was no sig-
`nificant difference between the likelihood of observing a change
`in FFR classification in any of the 7 different pressure responses.
`
`Discussion
`This is the largest study to investigate the effect of intravenous
`adenosine on the pressure recordings used to calculate FFR
`and the first to link changes in peripheral resistance and com-
`pliance with changes in the coronary perfusion pressure and
`measurement of FFR at different measurement time points.
`Our results are consistent with pressure changes observed
`in other studies, which were focused mainly on changes in
`coronary artery pressure.8–11 The fall in Pd pressure after
`intravenous adenosine is attributable to a combination of a
`fall in aortic perfusion pressure, in addition to microcircu-
`latory vasodilatation. Using Pr analysis, a close relationship
`was found between the changes in Pa pressure and changes in
`large artery compliance and resistance. This study shows how
`changes in systemic blood pressure (BP) at peak and stable
`hyperemia may directly alter the FFR value, and how this
`could lead to differences in lesion classification irrespective
`of whether either the FFR ≤0.80 or <0.75 treatment thresh-
`olds are used.
`
`Downloaded from
`
`http://circinterventions.ahajournals.org/
`
` by guest on October 4, 2017
`
`CATHWORKS EXHIBIT 1012
`Page 658 of 662
`
`

`

`Tarkin et al
`
` Hemodynamic Response to Intravenous Adenosine
`
` 659
`
`Figure 5. Overall change in aortic pressure (Pa) and distal intracoronary pressure (Pd), and respective ratios and gradients, at baseline,
`peak, and stable hyperemic conditions. Mean Pa and Pd were calculated for the entire population, and for each of the different classifi-
`cation groups at baseline, peak, and stable hyperemic conditions. After this, the mean pressure gradient and pressures ratios were cal-
`culated. At peak hyperemia, Pd falls disproportionately to Pa, resulting in a fractional flow reserve (FFR) value that crosses the FFR 0.80
`clinical treatment threshold, and which is lower than the FFR value when calculated at stable hyperemia. Similar fluctuations are seen
`when using the pressure gradient (Pa−Pd) or pressure ratio (Pd/Pa).
`
`Hemodynamic Responses to Intravenous Adenosine
`After intravenous adenosine infusion, we observed 7 different
`patterns of Pa and Pd. The pattern of Pr change, as a measure
`of peripheral resistance and compliance, mirrored Pa in each
`
`of the 7 groups, which suggests that the mechanism underlying
`the hemodynamic response to intravenous adenosine is largely
`mediated by changes in large artery compliance and resis-
`tance. A brief initial rise in systemic BP was often observed
`in the first moments after adenosine infusion. This may result
`
`Table 2. Logistic Regression Analysis of Clinical Characteristics in Relation to Observed Clinical
`Categorization Change in FFR Between Peak and Stable Hyperemia
`
`FFR Threshold
`
`Odds Ratio
`
`Age
`Sex
`Heart rate
`% Luminal stenosis
`Hypertension
`Hyperlipidemia
`Diabetes mellitus
`Smoking
`Chronic renal failure
`Previous MI
`Troponin positive
`Multivessel disease
`Previous CABG
`
`1.00
`0.74
`0.99
`1.00
`0.72
`0.44
`1.07
`0.91
`1.01
`0.44
`1.76
`0.90
`1.23
`
`≤0.80
`
`95% CI
`
`0.96–1.04
`0.30–1.84
`0.95–1.02
`0.98–1.03
`0.33–1.59
`0.19–1.00
`0.47–2.43
`0.41–2.03
`0.22–4.59
`0.06–3.39
`0.76–4.07
`0.40–2.05
`0.15–10.23
`
`P Value
`
`Odds Ratio
`
`0.99
`0.51
`0.54
`0.76
`0.42
`0.05
`0.87
`0.81
`0.99
`0.43
`0.19
`0.80
`0.85
`
`1.00
`1.68
`0.97
`1.00
`0.76
`1.48
`0.13
`1.98
`0.90
`1.75
`1.68
`1.02
`2.29
`
`≤0.75
`
`95% CI
`
`0.95–1.05
`0.37–7.67
`0.92–1.03
`0.96–1.04
`0.27–2.17
`0.40–5.46
`0.02–1.00
`0.68–5.72
`0.11–7.24
`0.37–8.28
`0.56–5.02
`0.35–2.96
`0.27–19.57
`
`P Value
`
`0.99
`0.50
`0.33
`0.98
`0.61
`0.56
`0.05
`0.21
`0.92
`0.48
`0.35
`0.97
`0.45
`
`CABG indicates coronary artery bypass grafting; CI, confidence interval; FFR, fractional flow reserve; and MI, myocardial infarction.
`
`Downloaded from
`
`http://circinterventions.ahajournals.org/
`
` by guest on October 4, 2017
`
`CATHWORKS EXHIBIT 1012
`Page 659 of 662
`
`

`

`660
`
` Circ Cardiovasc Interv
`
` December 2013
`
`from the passage of adenosine through the pulmonary circu-
`lation, inducing a short reflex peripheral vasoconstriction.12
`After this reflex response, adenosine acts to decrease periph-
`eral resistance with an associated fall in systemic BP.
`Pressure, under conditions of stable hyperemia, is considered
`to be proportional to blood flow.13 However, our findings sug-
`gest that with central intravenous adenosine infusion, there is
`often an initial reflex rise in flow, seen as a rise in Pa. If FFR
`is measured prior to achieving stable hyperemia, the initial
`rise in Pa together with a large reduction in the Pd leads to the
`potential for overestimation of lesion severity. In contrast, under
`stable hyperemia, the FFR value is frequently higher, most
`often reflecting a fall in Pa and a rise in Pd, compared with
`peak measurements (Figure 5). It has been hypothesized that
`this increase in Pd in the face of continued falling Pa perfusion
`pressure reflects an autonomic vasoconstrictor mechanism of
`the coronary microcirculation. This has been termed paradoxi-
`cal microcirculatory vasoconstriction.14 In such cases, the coro-
`nary microcirculation behaves like any other vascular bed—by
`vasoconstricting under conditions of low perfusion pressure to
`increase distal intracoronary pressure, and preserve flow.
`Although adenosine is used to mimic exercise, recent stud-
`ies have demonstrated that this pharmacological response dif-
`fers from that during resistive nonpharmacological exercise.15
`During exercise, both peripheral and microcirculatory resis-
`tance falls, and perfusion pressure to the coronary and sys-
`temic circulation is preserved by an increased cardiac output,
`most evident through a widening of pulse pressure. However,
`during adenosine infusion, Pa falls because of the reduction in
`peripheral resistance.
`Over and above these differential responses between
`pharmacological and exercise vasodilatation, Siebes et al14
`described the influence of pressure on FFR measurements in
`a theoretical model. These findings suggest that differential
`responses in pressure may be more widespread and unpredict-
`able than frequently recognized. Our study supports the find-
`ings of this theoretical model and demonstrates that across the
`range of physiological stenoses routinely studied in assess-
`ment of intermediate stenoses, Pa, rather than reduction in
`coronary microcirculatory resistance, is on average respon-
`sible for the majority of the fall in Pd. This effect is most evi-
`dent with larger falls in Pa and least with smaller reduction
`in Pa. Furthermore, when a reduction in Pa is large, the per-
`ceived reduction in Pd/Pa calculation is not due to a worsen-
`ing of the stenosis, but due to lower Pa and Pd numbers. For
`example, if the pressure gradient is maintained constant, but
`Pa and Pd pressure lowered the pressure ratio (Pd/Pa) will get
`lower. This gives the artificial appearance that the stenosis
`has increased in physiological significance, but in fact this is
`not the case, and it is due to simple mathematics of ratio cal-
`culations. This suggests that where cardiac output cannot be
`increased to overcome the vasodilator-mediated reduction in
`peripheral pressure, FFR may be liable to be influenced by
`reduced left ventricular functional reserve.
`We observed differences in the response of Pa and Pd when
`measured in different vessels within the same patients, sug-
`gesting that the systemic responses to intravenous adenosine
`are variable not only between individuals, but also between
`measurements within the same patient. Such differences
`
`are regularly observed with other drugs, which are usually
`expected to behave differentially within and between patients,
`frequently having different doses and different actions in differ-
`ent vascular beds in the same individual. It is, thus, likely and
`not at all surprising to see similar intra- and interpatient differ-
`ential responses with adenosine.16 These findings may in part
`help explain the intrapatient variability of FFR classification.
`
`Impact of Changes in Systemic BP on FFR
`In this study, we found that changes in systemic BP caused by
`intravenous adenosine can lead to alterations in FFR lesion clas-
`sification potentially affecting clinical management decisions.
`Specifically, we observed in 1 in 11 (9%) cases, a difference in
`classification when measurements were made at peak and stable
`hyperemic pressures using the widely acknowledged 0.8 FFR
`treatment threshold, and observed a difference in 1 in 19 (5.2%)
`cases using 0.75 FFR treatment threshold. Furthermore, mean
`peak and stable FFR values for the entire study group similarly
`crossed below and above the 0.8 FFR threshold. These findings
`highlight the importance of measuring FFR once stable hyper-
`emic pressures are achieved as was the methodology followed
`in the FAME and DEFER studies.3–5 Otherwise, this could lead
`to significant alterations in FFR classification. Ideally, measure-
`ments should only be made when stable hyperemia is achieved
`after ≥60 seconds of stable intravenous adenosine infusion.
`Using intravenous adenosine via a central venous line main-
`tains a steady state, stabilizing hemodynamic conditions, and
`creates the optimal conditions for physiological lesion assess-
`ment. However, even when using the methodology used in the
`large clinical trials from which the outcome data were gener-
`ated, differences between peak and stable measurement occurs.
`Although theses difference may seem small whenever classifica-
`tion and therefore treatment strategy is changed between the ini-
`tial (numerically lower) FFR measurement and that achieved at
`stable hyperemia, it will mean that unintentionally clinical deci-
`sion making (to stent or defer therapy) will be made on a basis
`different from that used in the studies in which clinical outcome
`data have been generated. Importantly, all of these patients fall
`within the clinically important intermediate FFR 0.6 to 0.9 zone,
`and the impact of this difference in clinical decision making on
`clinical outcomes has not been rigorously assessed.
`
`Study Limitations
`This is the first study to our knowledge to demonstrate the
`impact of systemic hemodynamic change during measure-
`ment of FFR, and further studies are required for validation.
`This was a retrospective analysis of patients undergoing FFR
`measurements for assessment of intermediate coronary steno-
`ses (mean stenosis severity, 49%; FFR, 0.83). It is, therefore,
`representative of a typical clinical workload where assessment
`of intermediate stenoses is in question, and typical clinical
`practice. Our study focused on hemodynamic changes attrib-
`utable to intravenous adenosine infusion; however, some cli-
`nicians use bolus intracoronary adenosine. Measurement of
`FFR using intracoronary adenosine has been validated by
`large clinical outcome studies3,17 and may have less impact on
`systemic BP than intravenous adenosine.10 It is possible that
`differences in classification of FFR between intracoronary and
`intravenous adenosine may be attributable to the findings of
`
`Downloaded from
`
`http://circinterventions.ahajournals.org/
`
` by guest on October 4, 2017
`
`CATHWORKS EXHIBIT 1012
`Page 660 of 662
`
`

`

`Tarkin et al
`
` Hemodynamic Response to Intravenous Adenosine
`
` 661
`
`our study. Future studies, which include systematic measure-
`ment of both intravenous and intracoronary adenosine admin-
`istration, would be needed to confirm this.
`
`Conclusions
`Intravenous adenosine significantly affects pressures in both
`systemic and coronary vascular beds. This can lead to signifi-
`cant differences in diagnostic classification when measure-
`ments are made between peak and stable FFR. To maximize
`the likelihood of obtaining the correct FFR classification, mea-
`surements should be made under conditions of stable hyper-
`emia, as performed in the landmark FFR outcome studies.
`
`Sources of Funding
`This work was supported by the National Institute for Health Research
`(NIHR) Imperial College Biomedical Research Centre. Drs Davies
`(FS/05/006), Francis (FS/10/038) and Petraco (FS/11/46/28861) are
`British Heart Foundation fellows. Drs Sen (G1000357) and Nijjer
`(G1100443) are Medical Research Council fellows.
`
`Disclosures
`Dr Davies is a consultant for Volcano Corporation. The other authors
`report no conflicts.
`
`References
` 1. Levine GN, Bates ER, Blankenship JC, Bailey SR, Bittl JA, Cercek B,
`Chambers CE, Ellis SG, Guyton RA, Hollenberg SM, Khot UN, Lange
`RA, Mauri L, Mehran R, Moussa ID, Mukherjee D, Nallamothu BK,
`Ting HH; American College of Cardiology Foundation; American Heart
`Association Task Force on Practice Guidelines; Society for Cardiovascular
`Angiography and Interventions. 2011 ACCF/AHA/SCAI Guideline for
`Percutaneous Coronary Intervention. A report of the American College
`of Cardiology Foundation/American Heart Association Task Force on
`Practice Guidelines and the Society for Cardiovascular Angiography and
`Interventions. J Am Coll Cardiol. 2011;58:e44–e122.
` 2. Hamm CW, Bassand JP, Agewall S, Bax J, Boersma E, Bueno H, Caso P,
`Dudek D, Gielen S, Huber K, Ohman M, Petrie MC, Sonntag F, Uva MS,
`Storey RF, Wijns W, Zahger D; ESC Committee for Practice Guidelines.
`Guidelines on myoc