430 
SURGERY AND TRANSPLANTATION 
ences between the normal and denervated left 
ventricular response to exercise in dogs by si- 
multaneous measurements of ventricular pres- 
sure, internal transverse diameter, and aortic 
flow. 
METHODS 
Seventeen American foxhounds, weighing 15 
to 23 Kg, were trained to run freely on a mo- 
tor-driven treadmill. After successful training, 
a left thoractomy was performed under sterile 
conditions. A miniature solid-state pressure 
transducer (7 mm diameter) was inserted into 
the left ventricular cavity through a small inci- 
sion at the apex, and fixed to the epicardium by 
a pursestring suture. The gauge was positioned 
approximately 8 millimeters inside the endo- 
cardial surface to avoid muscular entrapment 
during systole. A snug-fitting electromagnetic 
flow transducer was positioned around the as- 
cending aorta over a thin layer of Dacron mesh 
which effectively prevented aortic rupture due 
to erosion, but did not significantly alTect cali- 
bration. 
For measurement of internal left ventricular 
diameter two 5 millimeter discs of lead titan- 
ate-zirconate piezoelectric crystal, mounted in 
lucite housings and capped with convex lucite 
lenses, were positioned on opposite endocardial 
surfaces of the left ventricle across the maxi- 
mum transverse diameter in an anteroposterior 
orientation. Proper placement of the transduc- 
ers was determined by palpation with the oper- 
ator's index finger inserted through the left 
atrial appendage and mitral valve. After palpa- 
tion, a 14 inch length of PE 240 tubing, through 
whose terminal portion a 6 inch, 17 gauge nee- 
dle had been inserted, was passed through the 
ventricle across the selected diameter. Follow- 
ing transfixion of the ventricle, the needle was 
withdrawn and a loop of PE tubing was pulled 
through the mitral valve and exteriorized via 
the left atrial appendage. The loop was then cut 
and the terminal portions of the transducer 
leads were inserted into the open ends of the PE 
tubing. By withdrawing the other ends of the 
PE tubing from the ventricular surfaces the 
transducers were pulled into the ventricle and 
positioned against the endocardium. Proper 
placement and avoidance of chordal structures 
were again assured by palpation within the ven- 
tricle. The left atrial appendage was then li- 
gated. Inflow occlusion was not employed, and 
blood loss was negligible. 
At the conclusion of operation, all lead wires 
were passed subcutaneously from the chest inci- 
sion to the back of the neck where they were ex- 
teriorized and bandaged. In 11 of the 17 dogs 
regional cardiac denervation was performed 
immediately prior to the implantation of instru- 
ments. There was no significant difference in 
body weights between normal and denervated 
dogs (20.5 vs. 21.7 Kg). Denervation was ac- 
complished essentially as described by Cooper'' 
except that the caudal third of pericardium was 
not removed and the inferior pulmonary veins 
were not denuded. Criteria for verification of 
denervation included absence of sinus arrhyth- 
mia, lack of a heart rate startle response, ab- 
sence of heart rate changes with atropinization 
(2 mg) or administration of tyramine (.06 
fjLg/Kg) , and in three animals, absence of heart 
rate or contractile responses to supramaximal 
stimulation of the cervical vagi and stellate 
ganglia under pentobarbital anesthesia. 
All animals were treated with penicillin and 
streptomycin post-operatively, and were subse- 
quently studied 10 days to four weeks after op- 
eration, following clinical recovery. Daily rectal 
temperatures were also recorded early post- 
operatively, and, in the presence of fever, Chloro- 
mycetin was also given for 5-7 days. 
Following recovery, animals were exercised 
on a treadmill at least three times at daily inter- 
vals before study. At the time of study, a stand- 
ard five minute run on the treadmill at 3.5 mph 
and 15° grade was performed. All animals eas- 
ily completed this workload which was chosen 
to represent light exercise.'' ® Control measure- 
ments were obtained with dogs standing on the 
treadmill immediately before the start of the 
run. Interval records during the run were ob- 
tained at 30 seconds, one minute, and at one 
minute periods thereafter for the duration of 
the run. Additional recordings were made at 
one and five minutes during recovery. Signals 
were recorded on a Brush 480 recorder at paper 
speeds of 25 and 100, and in some cases 200 
mm./sec, and also on magnetic tape (Ampex 
