JACC Vol. 28, No.

5
November I. 1996:1432-5

LETTERS TO THE EDITOR

careful retrograde mapping of the atrial exit site of the fast pathway
before radiofrequency ablation can help avoid inadvertent heart block
in these patients.
BRUCE B. LERMAN, MD
KENNETH M. STEIN, MD
STEVEN M. MARKOWITZ, MD
The New York Hospital-Comell Medical Cemer
525 East 68 Street
New York, New York 10021
References
1. SungRJ. YoungC. LiemLB. SanathanaMurthy MG. Localizationof retrogradefast and
slowatrioventricularnode pathwayexitwith inductionof atypicalventriculoatrialWcnckebach periodicity[abstract]. Circulation1992:86Suppl 1:I-202.
2. EnglesteinED. SteinKM.MarkowitzSM.LermanBB.Posteriorfastatrklvenlricularnode
pathways: implicationsfor radiofi'equencycathcler ablation of atrioventricular node
reentrant tachycardia. J Am CoilCardiol 1996:27:1098 105.
3. Wilber DJ, Kopp DP. Olshansky B, Kall JG. Lippman N. Lerman BB. Spectrum of
atrioventricularnodal reentry[abstract]. J Am CoilCardiol 19~)3:2l:281A.
a. Lauer MR, Young C, MunsifA. Sung RJ. Identificationof patients at high risk of
developingcompleteatrioventricularblockduring ablation nf slowpathwayconduction
[abstract]. Circulation1993:88Suppl I:1-203.
5. WilberDJ, KallJG. OlshanskyB, KoppD, LippmanN, LermanBB. Fast-slowatrioven
tricular nodal reentry:anteriorversusposteriorretrogradepathways[abstract]. J Am Coil
Cardiol 1993:21:355A.

Supplemental Oxygen Administration and

Congestive Heart Failure
We read with concern the interesting work of Haque et al. (1), and we
wish to raise two important points.
1. Should we be surprised by a diminution of cardiac output (CO)
seconda W to supplemental oxygen administration? It seems sensible
that the increase in arterial oxygen content (Cao:) obtained in this way
could consequently permit a decrease in CO to maintain constant
oxygen delivery,"(Doe) at a given work load. Moore et al. (2) described
such a variation in CO secondary, to an increased inspired oxygen
concentration in patients with chronic congestive heart failure undergoing exercice tests. Interestingly, the decrease in CO observed in their
case was associated with reduced subjective scores of fatigue and
breathlessness. It would have been appropriate for Haque et al. (1) to
provide more information about the evolution of the clinical status of
their patients during oxygen administration. In other words, we have to
be sure that the reported observations truly reflect the detrimental
effects of oxygen therapy. The formula of Do e permits prediction of
the theoretic variation in CO that should be expected in response to a
given increase in Ca%. For example, in the case of the patients of
Haque et al., and assuming a normal hemoglobin concentration of
15 g/dl, it can be calculated that the expected decrease in CO would be
8.92% in experiment 1 and 8.16% in experiment 2. Undoubtedly, the
reported decrease in CO remains greater (16.22% and 23.68%,
respectively), but the significance of the results should be discussed in
light of this fundamental notion.
2. It can be calculated from the data of Haque et al. that oxygen
consumption (Vo2) was not constant during the investigation. In fact,
Vo 2 values varied from 370.7 ml/min at 21% fraction of inspired
oxygen (Flo2) to 317.8 ml/min at 100r~ FI% in experiment 1 (Vo 2
(21%) - V% (100%)/Vo: (100%) = 14.27%}, and from 354.2 ml/min
at 21% Flo e to 281.3 ml/min at 100% Flo~ in experiment 2 (V% (21%)
- Vo e (100%)/Vo e (100%) = 20.58%). Even if these calculated Vo e

Downloaded From: https://content.onlinejacc.org/ on 07/19/2016

1433

values are the result of a formula that includes CO, which implies a
certain amount of mathematical coupling between Vo 2 and oxygen
deliveU (Do2) calculated values, there is no guarantee that an external
factor did not modi~ Vo2 and, secondly, Do, and CO. It has been
previously hypothesized (3) that hyperoxia-related vasoconstriction
could decrease V02 by increasing peripheral shunting, which could
explain the increased mixed venous oxygen saturation (Sv02) and
decreased oxygen extraction ratios (ERo2) reported in the study of
Haque et al. (1). Another explanation could be that their patients were
already on the oxygen supply/V% dependency slope of their VOe/DO2
relations. This could explain the fact that V% decreased in response to
any reduction of CO and D%. Some factors support this hypothesis,
such as the extremely poor ejection fraction of the study patients or the
relatively low Svo2 measured. In contrast, the values of Do 2 calculated
in both experiments (Do: (21~) = 386.86 ml/min per m 2 and Do e
(100%) = 355.64 ml/min per m-" in experiment I and Do e (21%) =
401.56 ml/min per m 2 and D% (100%) = 333.5 ml/min per m 2 in
experiment 2, assuming a mean surface body area of 1.8 m-') remain
greater than the previously described Do~ critical limit of 330 ml/min
per m 2 (4) below which oxygen supply/Voa dependency appears. The
same can be said for ER% (ER% (21%) = 53.23% and ER%
(100%) = 49.65% in experiment 1 and ERo2 (21%) = 49.00% and
ER% (100%) = 46.86% in experiment 2), which remains lower than
the reported critical ERo2 limit of 70q (5). The latter two argue for a
o~gen supply/Voe-independent situation. If Haque et al. (1) had
reported the lactate serum levels of their patients, the question of
oxygen supply?do 2 dependency or peripherical shunting would have
been avoided. Actually, lactate serum level usually increases in situations of o~gen supply/V% dependency, as well as in any form of Do,
insufficiency (6), and should have been used as the final tracer of
oxy,gen imbalance. Despite these remarks, Haque et al. (1) have
opened an important and exciting debate that merits further development even if the clinical implications of such a study are not obvious.
Indeed, in a patient with severe heart failure steady state, why should
oxygen therapy be initiated? In contrast, if the same patient develops
any form of Do e insufficiency, he or she will undoubtly take advantage
of supplemental oxygen administration.
SERGE M. BROKA, MD
ANNE R. DUCART, MD
EDITH L. COLLARD, MD
KURT L. JOUCKEN, MD
Department o[ Am'sthesiolog~'
UniversiO" Clinics UCL of Mont-Godhme
B-5530 Yvoir, Belgium

References
1. HaqueWA. BoehmerJ. ClemsonBS. LeuenbergerUA. SilberDH. SinowayLI. Hemodynamiceffectsof supplementalo~gcn administrationin congestiveheart failure. J Am
CoilCardio1199(*:27:353-7.
2. Moore DP, Weston AR. HughesJMB. OakleyCM. ClelandJGF. Effectsof increased
inspired oxygenconcentrationson exerciseperformancein chronicheart failure. Lancet
1992:339:850-3.
3. Reinhart K, Specht M. Fhhring U, Mayr O, EyriehK. Einfluder praoxygcnierungauI
hfimodynamikund sauersto~erbrauch. Anaesthesist1989:38:233-7.
4. ShibutaniK. KomatsuT, Kubai K. SarchalaV. Kumar V. Bizarri DV. Critical levelof
oxygendeliveU in anesthetizedman. Crit (2arcMeal1983:11:640-3.
5. SamselRW. NelsonDP. SandersWM.WoodLDH.SchumackcrPT. Effectsof endotoxin
on systemicand skeletalmuscleO, extraction.J Appl PhDiol 1988:65:1377-82.
O. HauptMT.GilbertEM.CarlsonRW.FluidloadingincreasesoxTgenconsumptionin septic
patients with lacticacidosis.Am Rev Rcspir Dis 1985:131:912-6.

1434

LETTERS T O T H E E D I T O R

Reply
We appreciate the keen interest of Broka et al. in our work on the
hemodynamic effects of oxygen. They raised two concerns. The first
relates to the importance and mechanism of our observation demonstrating a diminished cardiac output in response to supplemental
oxygen. The second relates to the calculated oxygen consumption
values in our study and the fact that oxygen consumption decreased as
supplemental oxygen was added.
With regard to the first concern, the majority of subjects did not
become symptomatic as we gave supplemental oxygen. However, one
subject became short of breath and diaphoretic in response to this
intervention. Another subject who was studied after completing this
project also became diaphoretic as oxygen was delivered. We did not
include these data in the original report because we did not systematically examine symptoms and considered these data to be too anecdotal. However, as we emphasized in our report, pulmonary capillary
wedge pressure increased, confirming the finding that oxygen had a
detrimental hemodynamic effect. We agree that measurements of
plasma lactate would have been useful. Unfortunately, at the time of
these studies, we did not think of performing these measurements.
What is the mechanism for the decrease in cardiac output? We
would expect a change in cardiac output to be associated with either a
change in loading conditions or a change in inotropy. As mentioned
above, pulmonary capillary wedge pressure increased in response to
oxygen, providing evidence that preload increased. This would suggest
that cardiac output diminished at a time when Starling forces would
dictate that cardiac output should increase. Therefore, one could only
conclude that there must be a change in either inotropy or systemic
resistance. Although we did not measure inotropy directly, a decrease
in inotropy is typically accompanied by evidence of reflex sympathoexcitation. We found no changes in heart rate or peroneal nerve
muscle sympathetic nervous system activity. Thus, we would surmise
that inotropy did not decrease. We did calculate system vascular
resistance and found it to increase fairly dramatically. Therefore, our
data support the contention that oxygen acted as a direct vasoconstrictor, thereby increasing peripheral vascular resistance.
We have further pursued the issue of changes in systemic resistance
in our experiments in patients with a left ventricular assist device. We
have performed studies in six such patients and found a consistent
relation between increased systemic resistance and the administration
of oxygen (unpublished observations). Further, in normal subjects, we
have performed experiments on forearm blood flow in response to
10 min of forearm ischemia. We have observed that peak vasodilation
is reduced with the administration of supplemental oxygen (1). The
exact mechanism for this change in systemic resistance is currently
unknown,
In terms of the insightful discussion on oxygen delivery of Broca et
al., we completely agree that the change in calculated oxygen consumption in our data would bring into question the dependency of oxygen
consumption on oxygen delivery. Ideally, one would measure oxygen
consumption during such experiments. Measuring oxygen consumption during oxygen delivery is fraught with a number of inaccuracies
and, accordingly, we did not pursue these measurements. The work
cited in their letter was performed in anesthetized patients with
coronary disease (2). These subjects differed considerably from those
in our study in that they were anesthetized, and there is no comment
on the presence or absence of congestive heart failure. These are
important considerations because spontaneous ventilation in the presence of increased filling pressures would substantially increase the
work of breathing and total oxygen consumption. This could substan-

Downloaded From: https://content.onlinejacc.org/ on 07/19/2016

JACC Vol. 28, No. 5

November 1, 1996:1432-5

tially alter the relation between oxygen consumption and oxygen

delivery so aptly described in your letter.
With regard to the estimated oxygen consumption values in the
letter of Broca et al., there were some assumptions made that would
overestimate the calculated oxygen delivery. In reviewing the data in
experiment 2, we found that oxygen delivery ranged from 333 ml/min
on room air to 270 ml/min with 100% oxygen. Accordingly, we were in
the previously described range where there is a dependency of oxygen
consumption on oxygen delivery. Additionally, from the data that
Broca et al. referenced, one would predict a 21% decrease in oxygen
consumption. Based on our patients, a 29% change was observed. As
mentioned above, because the patient groups and clinical conditions
are different, it would be inappropriate to directly apply the previous
data to our patients.
In conclusion, our work dealt primarily with the hemodynamic
effects of supplemental oxygen in conscious humans with severe heart
failure. We agree with Broca et al. that other "clinical" measures
during oxygen administration in patients with heart failure are warranted.
JOHN P. BOEHMER, MD
LAWRENCE L. SINOWAY, MD
The Pennsyh'ania State UniversiO'
The Milton S. Hershey Medical Center
P.O. Box 850
Hershey., Penn~lvania 17033

References
1. Crawford P, Good PA, Gutierrez E. Feinberg JH, Silber DH, Sinoway L1. The effects of
oxygen on forearm dilator capacity [abstract]. FASEB J 1996;10:A590.
2. Shibutani K, Komatsu T, Kubai K, Sarchala V, Kumar V, Bizarri DV. Critical level of
oxygen delivery, in anesthetized man. Crit Care Med 1983;11:640-3.

Selection Bias in Thrombolytie Trials

Jha et al. (1) compared the characteristics and mortality outcomes of
Canadian patients who participated in two thrombolytic trials
(GUSTO and LATE) with those in patients with acute myocardial
infarction who did not participate. Administrative discharge data were
used for nonparticipants to obtain demographic and "comorbidity
scores." No physiologic data were available for nonparticipants. The
authors found that trial participants were younger, more likely to be
male and had a lower comorbidity score. Participants also had lower
in-hospital mortality "after adjustment for age, gender, revascularization and comorbidity scores." The authors conclude that there is
selection bias in the recruitment of trial participants, with favoring of
lower risk patients.
In essence, what the authors have described are differences in
demographics, comorbidity scores and mortality between a population
of patients who are eligible for thrombolysis (participants) compared
with an unselected group of all patients with acute myocardial infarction (nonparticipants). Nonparticipants thus include both thrombolysis
eligible and ineligible patients. It is well known that many patients with
acute myocardial infarction are ineligible for thrombolysis on the basis
of late presentation, absence of chest pain, nondiagnostic electrocardiographic findings, uncontrolled hypertension or other contraindications (2). Data reveal that these patients are older, more likely to be
male and more likely to have a higher degree of comorbidity (3).