General Cardiology – Basic Science
Cardiology 2000;93:234–241
Temperature-Controlled High Frequency
Ablation for Creation of Transmyocardial
Channels: In vivo Validation of a Novel Method
Ulrich Dietz a Mike Otto b Michael Buerke a Olaf Eick d Reza El Odhi a
Alp Förderer a Gerd Rippin c C. James Kirkpatrick b Jürgen Meyer a
Harald Darius a
a Medical Clinic II, b Institute of Pathology and c Institute for Medical Statistics and Documentation,
University Hospital, Mainz, Germany; d Medtronic, Bakken Research Center, Maastricht, The Netherlands
Key Words
Transmyocardial revascularization W High frequency
ablation
Abstract
Objective: We investigated the feasibility and short-term
effects of a novel procedure to create intramyocardial
channels by means of high frequency (HF) ablation in a
rabbit in vivo model. Methods: A flexible catheter ending
in a cylindrical electrode (diameter 0.7 mm) with a sharp-
ened tip was used for HF energy application following
transmyocardial insertion. Power-controlled or energy-
controlled energy applications were performed in 16
anesthetized rabbits after thoracotomy with a follow-up
for 3 h. Assessment of myocardial channels and the
necrotic zone was performed by morphometric quantifi-
cation in serial sections. The ferret diameter was used to
compare channel dimensions and the extent of necrosis.
Results: Thirty-nine power-controlled and 54 tempera-
ture-controlled HF applications were performed. The
shape of identified channels was round in 71% and 69%
This study was presented in part at the 71st Scientific Sessions of the
American Heart Association 1998 and the 20th Congress of the Euro-
pean Society of Cardiology 1998.
© 2000 S. Karger AG, Basel
ABC 0008–6312/00/0934–0234$17.50/0
Fax + 41 61 306 12 34
E-Mail karger@karger.ch Accessible online at:
www.karger.com www.karger.com/journals/crd
Received: February 7, 2000
Accepted after revision: April 27, 2000
had a lumen patency of 62/3 of the channel. Ferret diam-
eter of the channels was 414 B 180 Ìm and of the necrot-
ic zone 3,558 B 1,200 Ìm. In temperature-controlled
applications, channel dimensions were strongly in-
fluenced by the maximum tissue temperature and the
duration of energy delivery (Tmax: p = 0.0006; duration:
p = 0.003). Channel and necrosis dimensions correlated
better with biometric parameters in temperature-con-
trolled compared with power-controlled applications.
Conclusion: Mechanically created transmyocardial chan-
nels can be stabilized by HF heating of the surrounding
tissue. A high percentage of these channels remain
patent. The channel dimensions are closely correlated
with maximum temperature and duration of energy de-
livery in a temperature-controlled application mode.
Copyright © 2000 S. Karger AG, Basel
Introduction
Direct myocardial revascularization procedures have
been proposed for patients who are not suitable for any
further interventional or surgical revascularization proce-
dure because of end-stage coronary heart disease [1]. Clin-
ical application of laser revascularization procedures
(transmyocardial and percutaneous myocardial revascu-
larization) resulted in a marked improvement of symp-
Ulrich Dietz, MD
Deutsche Klinik für Diagnostik
Aukammallee 33
D–65191 Wiesbaden (Germany)
Tel. +49 611 577 267, Fax +49 611 577 577, E-Mail kardio.dietz@dkd-wiesbaden.de
toms [2]. However, in histological analyses done in pa-
tients who died after the procedure, the majority of chan-
nels were occluded because of scar tissue obliterating the
channels [3, 4]. We investigated high frequency (HF) cur-
rent to be used for the creation of transmyocardial chan-
nels. The feasibility of HF energy application in trans-
myocardial insertions performed with a specially de-
signed probe and the influence of the energy delivery
mode on the histological outcome are focal points of the
present study.
Methods
HF Device
The HF probe (Medtronic Bakken Research Center, Maastricht,
The Netherlands) consists of a 4-french catheter with a cylindrical
ablation electrode, 0.7 mm in diameter, 5 mm in length, which ends
in a sharpened tip. A thermocouple is incorporated in the center of
the ablation electrode. The device was connected to a high frequency
generator (Atakr, Medtronic CardioRhythm, Minneapolis, Minn.,
USA) for high frequency energy delivery (475 kHz) by either a tem-
perature-controlled (TCM) or a power-controlled (PCM) mode of HF
energy delivery. Energy delivery was automatically aborted when the
impedance was less than 50 ø or exceeded 250 ø. The HF power, the
ablation temperature and the impedance were recorded during ener-
gy delivery.
Animals Procedure
Animals were cared for according to the Principles of Laboratory
Animal Care. The experimental protocol was reviewed by the institu-
tional animal rights’ committee.
Sixteen adult rabbits, weighing 3.8 B 0.3 kg, were anesthetized
with an intravenous injection of thiopental sodium. Animals were
intubated and mechanically ventilated while midsternal thoracoto-
my and pericardiotomy were performed. The ablation electrode was
inserted transmurally at the anterior wall. Five to seven insertions
following HF energy application were performed in each animal. HF
energy applications were performed by one of two different modali-
ties: (1) the PCM or (2) the TCM. In PCM applications, the power
output was fixed ranging from 2 to 30 W with incremental steps of
2 W. Applications utilizing the TCM were performed with a preset
target temperature between 55 and 90 ° C with incremental steps of
5 ° C. The duration of HF applications was limited to 3, 6 or 10 s in
both modalities.
Hemodynamic monitoring was continued for 3 h. Thereafter, a
bolus of thiopental was administered for euthanasia. The hearts were
excised and coronary arteries were perfused by 4% buffered formal-
dehyde. Pieces of the ventricular wall containing the transmyocardial
channels were excised. Fixation was completed in a bath of buffered
formaldehyde for 5 days at 4 ° C.
Tissue Preparations and Analysis
Serial sections, each 5 Ìm thick, of paraffin-embedded myocar-
dial pieces were stained with hematoxylin/eosin (HE) and fuchsin for
qualitative and quantitative analysis of the channels and necrotic
zones, respectively. Histomorphometry was done using a calibrated
HF Transmyocardial Revascularization
in vivo
system for digital image acquisition and semiautomatic area calcula-
tion (Image Tools 1.28, University of Texas Health Science Center,
San Antonio, Tex., USA). The largest axis length, the smallest axis
length and the ferret diameter, which is computed as the diameter of
a circle having the same area as the object [sqrt (4 ! area/pi)], of
channels and necroses were assessed.
Statistics
Data were expressed as mean value B standard deviation. The
statistical significance of differences between groups was determined
by the Wilcoxon test. A multivariate analysis was performed using a
linear regression model. The variation coefficient (VC) was calcu-
lated by dividing the standard deviation by the mean value. Spear-
man correlation coefficient was calculated to compare continuous
variables. A p value of ^ 0.05 was regarded as statistically signifi-
cant.
Results
Thirty-nine power-controlled applications were per-
formed in 8 rabbits and 54 temperature-controlled appli-
cations in another 8 rabbits. Energy application was
aborted prematurely during 9 applications because of an
impedance exceeding 250 ø. Bleeding after removal of
the HF device occurred in all applications and was
stopped by manual compression. One animal died 1 h
after the intervention because of heart failure due to a
large anterior wall myocardial infarction. This was caused
by an application next to the left anterior descending
artery resulting in a proximal occlusion of the vessel. In all
other animals, heart rate and systemic blood pressure
were unchanged with respect to preinterventional values.
Experiments were terminated after 205 B 35 min. Mac-
roscopically, the puncture holes at the epicardial surface
were surrounded by a circular necrotic area in 56% of
applications (fig. 1).
Biometric Data
Power-Controlled Applications. In the PCM, the mean
power output was 14 B 8 W with an applied energy
between 6 and 225 J and a mean of 82 J. The measured
maximum tissue temperature was 73 B 13 ° C. Impe-
dance was 169 B 33 ø.
In general, a higher power output resulted in a higher
tissue temperature, but the maximum tissue temperature
was not associated with the power output (r = 0.2) or the
energy delivered (r = 0.32). The slope of the temperature
curve as well as the maximum temperature reached var-
ied widely for an identical power output (fig. 2a). Six
applications resulted in a tissue temperature of less than
55 ° C.
Cardiology 2000;93:234–241 235
 Temperature-Controlled Applications. In the TCM,
mean maximum tissue temperature measured 69 B
10 ° C. The power output which was adjusted automatical-
ly by the generator was on average 13 B 6 W with a deliv-
ered energy between 9 and 192 J and a mean of 67 J.
Impedance was 185 B 30 ø. The preset temperature was
reached within 1.5 s in all instances, which was achieved
by a high initial power output. The power output to main-
tain the temperature was much lower but varied widely

Fig. 1. Epicardial site of rabbit myocardium after application of HF

energy delivery using the PCM with a power output of 12 W and
maximum tissue temperature of 73 ° C. In the center, a hole can be
recognized which is surrounded by a circular zone of denatured tis-
sue.

Fig. 2. a Recording of tissue temperature

response in 4 power-controlled applications,
all of which were performed with a power
output of 20 W. b The tissue temperature
and the corresponding power output of 3
temperature-controlled applications with a
preset temperature of 65 ° C are depicted.
Note that in the PCM the temperature re-
sponse differs, while the power output is
identical. In the TCM, initially a high power
output is adjusted automatically by the gen-
erator, but the power output necessary to
maintain the tissue temperature differs for
these 3 applications.

236 Cardiology 2000;93:234–241 Dietz/Otto/Buerke/Eick/El Odhi/Förderer/

Rippin/Kirkpatrick/Meyer/Darius
Fig. 3. Myocardial specimen containing a
cross section of a channel created by power-
controlled HF energy delivery with a maxi-
mum tissue temperature of 75 ° C and a du-
ration of 10 s. The central channel is sur-
rounded by a necrotic zone which is delin-
eated from intact myocardium by a hypereo-
sinophilic band (arrows). The channel has an
ovaloid shape and is obstructed by less than
one third by debris (*). Note that the vessels
located within the necrotic zone (arrow-
heads) show no morphological alterations.
HE. !560.

Fig. 4. Myocardial specimen containing a

cross section of a channel created by temper-
ature-controlled HF energy delivery with a
temperature of 75 ° C and a duration of 6 s.
The channel has an oval shape without any
obstructive material. The surrounding tissue
shows moderate hemorrhage. The necrotic
area is demarcated by a typical dense fuchsi-
nophilic band (arrows). Fuchsin. !280.

between the applications using the temperature-con-
trolled mode (fig. 2b).
Comparison of Energy Delivery Modes. The slope of
the tissue temperature response exhibited a greater varia-
tion in the PCM as compared to the TCM (VC 0.7 vs. 0.2,
p = 0.001).
Histological Findings
Analysis of channels was performed in 905 sections
stained with HE. Channels created by the HF device
could be identified in 81 of 93 applications. Cross sections
of channels allowing qualitative and quantitative analysis
were present in 62 applications. No quantitative assess-
ment was performed in 5 channels because they were cut
in a longitudinal direction. In 14 applications, channels
could not be clearly delineated or a procedure-related
cause of tissue discontinuity was not certain, and in 12
applications no channel was recognizable by histology.
The channel shape was round to oval in 71% or flat-
tened in 29% of sections (fig. 3). Channel lumen was
obstructed by ^1/3 in 16%, 11/3 but ^2/3 in 45%, more
than 2/3 in 31%, and without any obstruction in 8%.
Obstructive material consisted of fresh thrombus in 94%
and cellular debris in 61%.

HF Transmyocardial Revascularization Cardiology 2000;93:234–241 237

in vivo
Fig. 5. Scatter plots and regression of the fer-
ret diameter and the corresponding tempera-
ture-time product. A close relation is present
in temperature-controlled (a) but not in
power-controlled (b) applications.
Quantitative assessment of channel dimensions was
performed in sections of 32 TCM applications and in
those of 30 PCM energy deliveries. The overall channel
area was 0.22 B 0.17 mm2, the smallest channel axis
amounted to 290 B 140 Ìm and the largest to 960 B
380 Ìm. The ferret diameter ranged from 80 to 1,254 Ìm
with a mean of 371 B 144 Ìm for the PCM applications
and from 79 to 1,130 Ìm with a mean of 460 B 200 Ìm
for the TCM applications. Quantitative assessment of the
necrotic zone was done in 167 sections stained with fuch-
sin (fig. 4). The ferret diameter of the necrotic zone was
4,460 B 1,460 Ìm (1,960–7,170 Ìm) for PCM applica-
238 Cardiology 2000;93:234–241
tions and was 3,320 B 1,180 Ìm (1,485–5,920 Ìm) for
TCM applications. The area of the necrotic zone mea-
sured 11 B 6.8 mm2 for all measurements. The ferret
diameter of the channels was correlated with the round-
ness of the channel (r = 0.43, p = 0.0019). No other signifi-
cant correlations between binary qualitative and quanti-
tative histological parameters were found.
Influence of Biometric Parameters
The tissue temperature was higher in channels with a
round or oval geometry than in those which were flat (74
B 12 vs. 69 B 12 ° C, p = 0.001). Channel dimensions
Dietz/Otto/Buerke/Eick/El Odhi/Förderer/
Rippin/Kirkpatrick/Meyer/Darius
Fig. 6. The ferret diameter of the necroses
and the maximum tissue temperature are
closely related in the temperature-controlled
application (a) but not in the application
using the power-controlled energy delivery
mode (b).
correlated closest with the duration of a tissue temperature
between 60 and 70 ° C (largest axis, r = 0.56; smallest axis,
r = 0.63; ferret diameter, r = 0.62; p !0.0002 for all values).
Correlation was less for the duration of tissue temperature
between 50 and 60 ° C, or exceeding 70 ° C (data not
shown). The channel ferret diameter in TCM applications
correlated with the temperature-time product (r = 0.60, p =
0.003) and the maximum tissue temperature (r = 0.53, p =
0.0018), whereas channel dimensions created in PCM did
not correlate with these parameters (fig. 5).
The channel ferret diameter after PCM applications
correlated with the power output (r = 0.53, p = 0.0025),
HF Transmyocardial Revascularization
in vivo
but not with the energy delivered; no such correlation was
found for channels created by TCM applications.
Dimensions of the necrotic zone were correlated with
the tissue temperature (largest axis, r = 0.61; smallest axis,
r = 0.62; ferret diameter, r = 0.65; p = 0.0001 for all val-
ues) for values of TCM applications, which also correlated
with the duration of the tissue temperature exceeding
70 ° C (r = 0.58, r = 0.59 and r = 0.60, respectively, for all
p ! 0.0002), whereas necrosis dimensions of PCM appli-
cations were not correlated with these parameters (fig. 6).
Dimensions of the channels and necroses were not corre-
lated with other biometric parameters.
Cardiology 2000;93:234–241 239
 Linear regression analysis revealed that the maximum
tissue temperature [probability of estimate (PE) 6.3, stan-
dard error of the estimate (SE) 3.2–10.1, p = 0.0006] and
the duration of energy delivery (PE 29, SE 14.6–43.2, p =
0.0003) were independent predictors for the dimension
of the channel ferret diameter in TCM applications. For
values obtained in PCM applications, no dependency of
the ferret channel diameter on biometric parameters
could be demonstrated. The maximum tissue tempera-
ture was predictive for the ferret diameter of the necrotic
zone (PE 85, SE 53–117, p = 0.0001) for values of both
application modes. No independent influence on necro-
sis dimensions was found for the duration of the energy
delivery or for the other parameters by linear regression
analysis.
Discussion
Numerous efforts have been made to restore myocar-
dial blood supply by an alternative route to the coronary
artery system in patients with end-stage coronary heart
disease. In the early 1960s, needle acupuncture of the
myocardium to achieve a nutritive blood flow from the
left ventricular cavity was investigated [5, 6]. However,
due to the architecture of the ventricular myocardium the
channels occluded rapidly with the exception of channels
created by a needle with a 4-mm diameter [7]. Laser was
also instituted to create channels by vaporization of myo-
cardium [8, 9]. Although short-term patency could be doc-
umented for some of the channels, most investigations
failed to show long-term patency [10, 11].
We hypothesized that tissue modulation by generation
of a local thermal injury can stabilize mechanically
created intramyocardial channels and thereby prevent
early occlusion. The aim of our study was to evaluate
whether HF current can be used to achieve a predictable
tissue denaturation surrounding mechanically created
channels. HF current results in resistive heating of tissue
which is in immediate contact with the active electrode
where the current density is much higher compared to the
large indifferent electrode. The current density decreases
by the square of the distance, and the temperature de-
creases by the fourth of the distance. Thus, a rapid and
well-controlled heating of tissue adjacent to the ablation
electrode can be achieved. Irreversible damage to the
myocardium has been documented for HF energy appli-
cations, if the myocardium is heated up to more than
50 ° C [12, 13]. Both the power-controlled and the temper-
ature-controlled energy delivery modes have been suc-
240 Cardiology 2000;93:234–241
cessfully used for endocardial tissue ablation [14, 15]. It
was intended that the present investigation should answer
3 questions: (1) is it feasible to create transmyocardial
channels by HF energy delivery? (2) is it possible to iden-
tify the channels histologically?, and (3) which energy
delivery mode results in the best reproducible tissue
effect, to serve as a prerequisite for a possible clinical
application?
The feasibility of this procedure could be clearly dem-
onstrated in this series of experiments. An incomplete
transmyocardial placement of the catheter caused an
abortion of the energy delivery process automatically,
which was a rare incident. Global left ventricular function
was not impaired, despite the fact that a rather large myo-
cardial area was treated. In 1 animal an anterior wall
infarction developed due to an unintentional misplace-
ment of the device, which resulted in a thrombotic LAD
occlusion.
In this study, we were able to demonstrate convincing-
ly that mechanically generated channels stabilized by HF
heating of the surrounding tissue stay patent in a high per-
centage of cases at least for a short period of time. The
hypothesis that tissue denaturation is needed to prevent
channels from collapsing was supported by the finding
that with maximum tissue temperature below 55 ° C no
open channels could be identified. In all channels identi-
fied, some signs of severe cell destruction next to the chan-
nels were visible microscopically.
Histologically, channels differed with respect to their
cross-sectional geometry and surface. In order to be able
to compare channel dimensions, the ferret diameter was
calculated, which has been shown to be superior to other
parameters in the quantification of objects with irregular
contours, although being a virtual parameter [16]. To
determine the impact of the duration of tissue heating, the
product of energy delivery duration and mean tempera-
ture was calculated, thereby integrating the temperature
over time. This parameter exhibited the closest correla-
tion to the channel’s ferret diameter, suggesting that the
stability of the channels may depend on the duration of
exposure to a given temperature. This assumption was
supported by the results of a linear regression model.
In contrast to the results of the channel analysis, the
extension of the necrotic zone was correlated with the
maximum tissue temperature. There was no dependency
found for the duration of tissue heating. These results con-
firm previous findings of other investigators [12] and can
be most likely explained by heat convection to the vascu-
lature. Therefore, an extended period of heat application
would not enlarge the necrotic zone. No relation was
Dietz/Otto/Buerke/Eick/El Odhi/Förderer/
Rippin/Kirkpatrick/Meyer/Darius
found between the extent of the necrotic zone and the
channel dimensions, suggesting that the maintenance of
channels can be achieved even by a small necrotic rim
that is able to withstand the compressing shear forces of
the contracting myocardium.
The last question to be answered is related to the mode
of energy delivery. Surprisingly, the temperature response
in power-controlled applications varied widely despite an
identical power output. In temperature-controlled appli-
cations, the preset temperature was kept constant due to
the automatic adjustment of the power output. The most
likely reason for this variation may be differences in heat
convection in the immediate vicinity of larger vessels.
The duration of HF delivery required to achieve tissue
temperatures exceeding 55 ° C was shorter in TCM as
compared to PCM applications. This was due to an initial
higher power output automatically adjusted by the gener-
ator in the TCM. The lower correlation between the chan-
nel ferret diameter and the temperature time product in
PCM applications may in part be explained by these find-
ings. Thus, the temperature-controlled energy delivery
mode seems to be superior to the PCM with respect to the
reproducibility of the biometric data and the histological
outcome.
Conclusion
Transmyocardial channels can be created by insertions
with a 0.7-mm probe and denaturation of the surrounding
tissue by heating using HF energy applied by a newly
designed catheter. A reproducible tissue temperature can
be achieved in the TCM, which requires a thermocouple
within the tip of the probe. By these means, intramyocar-
dial channels can be created and a high percentage of
these channels remains patent. Both, the tissue tempera-
ture and the duration of heating determine the dimen-
sions of the resulting channels. The extent of the sur-
rounding necrosis is small and depends on the maximum
tissue temperature. Thus, reproducible transmyocardial
channels with a small predictable necrotic zone can be
achieved by using a temperature-controlled HF device.

References

1 Vincent JG, Bardos P, Kruse J, Maass D: End
stage coronary disease treated with the trans-
myocardial CO2 laser revascularization: A
chance for the ‘inoperable’ patient. Eur J Car-
diothorac Surg 1997;11:888–894.
2 Horvath KA, Cohn LH, Cooley DA, Crew JR,
Frazier OH, Griffith BP, Kadipasaoglu K,
Lansing A, Mannting F, March R, Mirhoseini
MR, Smith C: Transmyocardial laser revascu-
larization: Results of a multicenter trial with
transmyocardial laser revascularization used as
sole therapy for end-stage coronary artery dis-
ease. J Thorac Cardiovasc Surg 1997;113:645–
653.
3 Burkhoff D, Fisher PE, Apfelbaum M, Kohmo-
to T, DeRosa CM, Smith CR: Histologic ap-
pearance of transmyocardial laser channels af-
ter 4½ weeks. Ann Thorac Surg 1996;61:1532–
1534.
4 Gassler N, Wintzer HO, Stubbe HM, Wull-
brand A, Helmchen U: Transmyocardial laser
revascularization. Histological features in hu-
man nonresponder myocardium. Circulation
1997;95:371–375.
5 Sen PK, Udwadia TE, Kinare SG, Parulkar
GB: Transmyocardial acupuncture. J Thorac
Cardiovasc Surg 1965;50:181–189.
6 Pifarre R, Jasuja ML, Lynch RD, Neville WE:
Myocardial revascularization by transmyocar-
dial acupuncture. J Thorac Cardiovasc Surg
1969;58:424–431.
7 Khazei AH, Kime WP, Papadopoulos C, Cow-
ley RA: Myocardial canalization. A new meth-
od of myocardial revascularization. Ann Tho-
rac Surg 1968;6:163–171.
8 Mirhoseini M, Muckerheide M, Cayton MM:
Transventricular revascularisation by laser. La-
sers Surg Med 1982;2:187–198.
9 Okada M, Shimizu K, Ikuta H, Horii H, Naka-
mura K: A new method of myocardial revascu-
larization by laser. Thorac Cardiovasc Surg
1991;39:1–4.
10 Hardy RI, Bove KE, James FW, Kaplan S,
Goldman L: A histologic study of laser-induced
transmyocardial channels. Lasers Surg Med
1987;6:563–573.
11 Kohmoto T, Fisher PE, Gu A, Zhu SM, DeRo-
sa CM, Smith CR, Burkhoff D: Physiology, his-
tology, and 2-week morphology of acute trans-
myocardial channels made with a CO2 laser.
Ann Thorac Surg 1997;63:1275–1283.
12 Haines DE, Verow AF: Observations on elec-
trode-tissue interface temperature and effect
on electrical impedance during radiofrequency
ablation of ventricular myocardium. Circula-
tion 1990;82:1034–1038.
13 Wittkampf FHM, Hauer RNW, Robles de Me-
dina EO: Control of radiofrequency lesion size
by power regulation. Circulation 1989;80:962–
968.
14 Borggrefe M, Budde T, Podczeck A, Breithardt
G: High frequency alternating current ablation
of an acessory pathway in humans. J Am Coll
Cardiol 1987;10:576–582.
15 Calkins H, Prystowsky E, Carlson M, Klein LS,
Saul JP, Gillette P: Temperature monitoring
during radiofrequency catheter ablation proce-
dures using closed loop control. Atakr Multi-
center Investigators Group. Circulation 1994;
90:1279–1286.
16 Otto MJ, Löw O, Schneider A: The norminal
section thickness – Importance of their correla-
tion for morphometry. Exp Pathol 1991;42:
129–136.

HF Transmyocardial Revascularization Cardiology 2000;93:234–241 241

in vivo