C ORR ES POND ENCE

Correspondence charide, with low-molecular-weight heparin initiated in close

proximity to elective hip surgery (six hours postoperatively).

RUSSELL HULL, M.B., B.S.

GRAHAM PINEO, M.D.
University of Calgary
Calgary, AB T2N 2T9, Canada
jeanne.sheldon@crha-health.ab.ca

A Synthetic Pentasaccharide for the Prevention
of Deep-Vein Thrombosis
To the Editor: Turpie et al. (March 1 issue)1 report the find-
ings of a rigorously performed double-blind, randomized
trial comparing a synthetic pentasaccharide (Org31540/
SR90107A) and a low-molecular-weight heparin (enox-
aparin) in patients undergoing total hip replacement. The
patients were randomly assigned to receive one of five dai-
ly doses of the pentasaccharide beginning 6 hours postop-
eratively (range, 4 to 8) or enoxaparin beginning 12 to 24
hours postoperatively. Two studies 2,3 evaluated the effect
of prophylaxis with low-molecular-weight heparin adminis-
tered in close proximity to surgery, either within two hours
before surgery or six hours after surgery, in patients under-
going elective hip surgery and found that low-molecular-
weight heparin was more effective than oral anticoagulants.
In contrast, the effectiveness of therapy with low-molecular-
weight heparin begun 12 hours preoperatively4 or 12 to 24
hours postoperatively5 was similar rather than superior to
that of oral anticoagulants. The timing of prophylaxis is cru-
cial; antithrombotic prophylaxis administered in close prox-
imity to surgery is more effective than delayed prophylaxis.
The conclusion of Turpie et al. that, as compared with
low-molecular-weight heparin, the synthetic pentasaccharide
has the potential to improve the riskbenefit ratio for the
prevention of venous thromboembolism is sound. However,
in the light of the evidence that initiating antithrombotic
prophylaxis in close proximity to surgery is more effective
than delaying prophylaxis, it is apparent that the superiority
of the pentasaccharide also reflects its initiation in the early
postoperative period. Further trials are required to compare
the new antithrombotic treatments, such as the pentasac-
Editors note: Drs. Hull and Pineo have received grants-
in-aid from Dupont, Emisphere Technologies, Leo Pharma-
ceutical Products, and Pharmacia.
1. Turpie AGG, Gallus AS, Hoek JA. A synthetic pentasaccharide for the
prevention of deep-vein thrombosis after total hip replacement. N Engl J
Med 2001;344:619-25.
2. Hull RD, Pineo GF, Francis C, et al. Low-molecular-weight heparin
prophylaxis using dalteparin in close proximity to surgery vs warfarin in hip
arthroplasty patients: a double-blind, randomized comparison. Arch Intern
Med 2000;160:2199-207.
3. Francis CW, Pellegrini VD Jr, Totterman S, et al. Prevention of deep-
vein thrombosis after total hip arthroplasty: a comparison of warfarin and
dalteparin. J Bone Joint Surg Am 1997;79:1365-72.
4. Hamulyak K, Lensing AWA, van der Meer J, Smid WM, van Ooy A,
Hoek JA. Subcutaneous low-molecular weight heparin or oral anticoagu-
lants for the prevention of deep-vein thrombosis in elective hip and knee
replacement? Thromb Haemost 1995;74:1428-31.
5. Hull R, Raskob G, Pineo G, et al. A comparison of subcutaneous low-
molecular-weight heparin with warfarin sodium for prophylaxis against
deep-vein thrombosis after hip or knee implantation. N Engl J Med 1993;
329:1370-6.
To the Editor: In their doseresponse trial comparing
the synthetic pentasaccharide Org31540/SR90107A with
enoxaparin for the prevention of deep-vein thrombosis af-
ter total hip replacement, Turpie et al. mention epidural and
spinal anesthesia in passing in the Methods section, but they
fail to provide any safety data regarding the sites of major
bleeding. According to Table 4 of their article, 9 of 260 pa-
tients in the control group (3.5 percent) had major bleeding
complications, as compared with 30 of 673 patients in the
pentasaccharide group (4.5 percent). Since in the setting of
continuous epidural or spinal anesthesia, the administration
of low-molecular-weight heparin has been associated with

INSTRUCTIONS FOR LETTERS TO THE EDITOR

Letters to the Editor are considered for publication (subject to editing and abridgment) provided they do not contain material that has been
submitted or published elsewhere. Please note the following: Your letter must be typewritten and triple-spaced. Its text, not including
references, must not exceed 250 words if it is in reference to a recent Journal article, or 400 words in all other cases (please provide a word
count). It must have no more than five references and one figure or table. It must not be signed by any more than three authors. Letters
referring to a recent Journal article must be received within four weeks of its publication. Please include your full address, telephone number,
fax number, and e-mail address. You may send us your letter by standard mail, fax, or e-mail.
Our address: Letters to the Editor New England Journal of Medicine 10 Shattuck St. Boston, MA 02115
Our fax numbers: 617-739-9864 and 617-734-4457
Our e-mail address: letters@nejm.org
We cannot acknowledge receipt of your letter, but we will notify you when we have made a decision about publication. We are unable to
provide prepublication proofs. Financial associations or other possible conflicts of interest must be disclosed. Submission of a letter constitutes
permission for the Massachusetts Medical Society, its licensees, and its assignees to use it in the Journal s various print and electronic
publications and in collections, revisions, and any other form or medium.

N Engl J Med, Vol. 345, No. 4 July 26, 2001 www.nejm.org 291

The New England Journal of Medicine

Downloaded from nejm.org on September 10, 2017. For personal use only. No other uses without permission.
Copyright 2001 Massachusetts Medical Society. All rights reserved.
 The Ne w E n g l a nd Jo u r n a l o f Me d ic i ne
neuraxial hematomas and even permanent paraplegia,1-3
Turpie et al. should outline their protocols recommenda-
tions for removing the catheter (if one was inserted) and
report whether any such events were observed in the in-
tention-to-treat population.
LAURENCE LANDOW, M.D.
7620 Old Georgetown Rd.
Bethesda, MD 20814
laurence.landow@mindspring.com
1. Wysowski DK, Talarico L, Bacsanyi J, Botstein P. Spinal and epidural
hematoma and low-molecular-weight heparin. N Engl J Med 1998;338:
1774.
2. Horlocker TT, Wedel DJ. Neuraxial block and low-molecular-weight
heparin: balancing perioperative analgesia and thromboprophylaxis. Reg
Anesth Pain Med 1998;23:Suppl2:164-77.
3. Landow L, Bedford RA. Low-molecular weight heparin, spinal hemato-
mas, and the FDA: whats wrong with this picture? Reg Anesth Pain Med
1999;24:8-10.
The authors reply:
To the Editor: We agree with Hull and Pineo that much
remains to be learned about the optimal scheduling of an-
ticoagulants when they are used before or soon after major
surgery. Our trial cannot contribute to the answer, since it
was essentially a doseresponse study of the pentasaccha-
ride. Our comparison with the currently recommended reg-
imen of enoxaparin prophylaxis was exploratory and pre-
liminary to formal phase 3 comparisons.
The optimal time to begin thromboprophylaxis with low-
molecular-weight heparin preoperatively or postopera-
tively is still controversial.1 The attractive concept that
administering low-molecular-weight heparin in close prox-
imity to surgery may increase its efficacy is based on indirect
comparisons between approved regimens and nonapproved
regimens, with warfarin as the reference treatment.2 This
concept has not, however, been unequivocally proved, since
a dose of 30 mg of enoxaparin twice daily that was initi-
ated within 24 hours after hip replacement proved more
effective than warfarin.3
When two different drugs are compared, factors other
than the timing of their initial administration may contrib-
ute to differences in efficacy. Moreover, the optimal dura-
tion of prophylaxis is still debated.1 Several studies suggest
that the risk of venous thromboembolism after hip replace-
ment may begin during surgery but may persist for up to
two months. Thus, the overall efficacy of a regimen is un-
likely to reflect only the effect of the timing of the first dose,
and the superior efficacy of pentasaccharide over enoxaparin
is likely to be due to its selective inhibition of activated fac-
tor X, its rapid onset of action, and its long half-life, which
results in a complete, 24-hour antithrombotic effect.4,5
The reports of neuraxial hematomas referred to by Dr.
Landow led to strengthened precautions for the use of an-
ticoagulation in conjunction with epidural or spinal anes-
thesia. Our patients underwent randomization postopera-
tively and only after the removal of any epidural catheter.
Injections were prohibited within two hours after the remov-
al of the catheter. In this dose-ranging study, one of the
first randomized patients who received 6 mg of pentasac-
charide, which is more than twice the dose selected for fur-
ther trials, had a neuraxial hematoma after five attempts to
place an epidural catheter preoperatively were unsuccessful.
292 N Engl J Med, Vol. 345, No. 4 July 26, 2001 www.nejm.org
The New England Journal of Medicine
Downloaded from nejm.org on September 10, 2017. For personal use only. No other uses without permission.
Copyright 2001 Massachusetts Medical Society. All rights reserved.
Subsequent patients could enter the trial after receiving re-
gional anesthesia only if the puncture was clean, was made
on the first attempt, and was uncomplicated by bleeding.
No further neuraxial hematomas occurred.
ALEXANDER G.G. TURPIE, M.D.
McMaster University
Hamilton, ON L8L 2X2, Canada
turpiea@mcmaster.ca
ALEXANDER S. GALLUS, M.D.
Flinders Medical Centre
Adelaide 5001, Australia
JACOB A. HOEK, M.D.
SanofiSynthelabo Research
Malvern, PA 19355
Editors note: Drs. Turpie and Gallus have received grants
from and served as consultants to SanofiSynthelabo Re-
search, which, along with Organon, is the manufacturer of
the anticoagulant pentasaccharide.
1. Geerts WH, Heit JA, Clagett GP, et al. Prevention of venous throm-
boembolism. Chest 2001;119:Suppl:132S-175S.
2. Hull RD, Pineo GF, Francis C, et al. Low-molecular-weight heparin
prophylaxis using dalteparin in close proximity to surgery vs warfarin in hip
arthroplasty patients: a double-blind, randomized comparison. Arch Intern
Med 2000;160:2199-207.
3. Colwell CW Jr, Collis DK, Paulson R, et al. Comparison of enoxaparin
and warfarin for the prevention of venous thromboembolic disease after to-
tal hip arthroplasty: evaluation during hospitalization and three months af-
ter discharge. J Bone Joint Surg Am 1999;81:932-40.
4. Turpie AGG, Gallus AS, Hoek JA. A synthetic pentasaccharide for the
prevention of deep-vein thrombosis after total hip replacement. N Engl J
Med 2001;344:619-25.
5. Rosenberg RD. Redesigning heparin. N Engl J Med 2001;344:673-5.
Low-Molecular-Weight Heparin in Patients
with Deep-Vein Thrombosis
To the Editor: The clinical importance of rapidly achiev-
ing a therapeutic activated partial-thromboplastin time with
respect to the treatment of venous thromboembolism with
unfractionated heparin has been underscored in various
studies1 and reviews.2 Breddin et al. (March 1 issue)3 com-
pared intravenous unfractionated heparin with subcutane-
ous weight-adjusted reviparin, given once or twice a day, as
a therapy for deep-vein thrombosis. Reviparin (a low-molec-
ular-weight heparin) was more effective than unfractionated
heparin in reducing the size of the thrombus, and twice-
daily administration of reviparin prevented recurrent throm-
boembolism better than did treatment with unfractionated
heparin. The patients received fixed initial doses of unfrac-
tionated heparin, with the doses adjusted according to daily
measurements of the activated partial-thromboplastin time.
This approach led to an unacceptably high number of pa-
tients with a subtherapeutic activated partial-thromboplas-
tin time after 48 hours (33 percent). A regimen involving
doses of heparin adjusted for the patients weight, as pro-
posed by Raschke et al.,4 would probably have led more
rapidly to therapeutic heparin levels and thus to fewer recur-
rent thromboembolic events. The statement that reviparin
regimens are more effective than a regimen of unfraction-
 CORR ES POND ENCE
ated heparin can be considered true only in the context of
a suboptimal heparin regimen.
HANS STRICKER, M.D.
GIORGIO MOMBELLI, M.D.
Regional Hospital
CH-6600 Locarno, Switzerland
hans.stricker@eoc.ch
1. Hull RD, Raskob GE, Brant RF, Pineo GF, Valentine KA. Relation be-
tween the time to achieve the lower limit of the APTT therapeutic range
and recurrent venous thromboembolism during heparin treatment for deep
vein thrombosis. Arch Intern Med 1997;157:2562-8.
2. Hirsh J. Heparin. N Engl J Med 1991;324:1565-74.
3. Breddin HK, Hach-Wunderle V, Nakov R, Kakkar VV. Effects of a low-
molecular-weight heparin on thrombus regression and recurrent throm-
boembolism in patients with deep-vein thrombosis. N Engl J Med 2001;
344:626-31.
4. Raschke RA, Reilly BM, Guidry JR, Fontana JR, Srinivas S. The weight-
based heparin dosing nomogram compared with a standard care nomo-
gram: a randomized controlled trial. Ann Intern Med 1993;119:874-81.
To the Editor: We would like to ask Breddin et al. to specify
a detail of the treatment protocol. Reviparin was given either
once or twice daily, at doses adjusted for body weight. Does
this mean that the twice-daily group actually got twice the
total daily dose received by the once-daily group, or was the
total dose given in a 24-hour period the same that is,
the twice-daily group received half the dose per injection?
CHRISTOPH PECHLANER, M.D.
WALTER GRITSCH, M.D.
Innsbruck University Hospital
A-6020 Innsbruck, Austria
christoph.pechlaner@uibk.ac.at
The authors reply:
To the Editor: Regarding the comments of Pechlaner and
Gritsch: we stated that patients in both the reviparin groups
received the same daily dose of reviparin, given either in
two divided subcutaneous injections every 12 hours or as a
single subcutaneous injection every 24 hours.
Regarding the comments of Stricker and Mombelli: we
would like to point out that conventional practice is to ad-
minister unfractionated heparin by intravenous infusion,
with the dose adjusted according to daily measurements
of the activated partial-thromboplastin time to achieve a
value 1.5 to 2.5 times the base-line level. To our knowledge,
a randomized, controlled trial has never been performed
comparing the efficacy of a heparin infusion adjusted for
the activated partial-thromboplastin time with a weight-
adjusted dose with the use of an objective method of as-
sessing thrombus regression and recurrent thromboembolic
events. Therefore, it remains speculative whether a weight-
adjusted regimen of unfractionated heparin would be more
effective in increasing thrombus regression and thus re-
ducing the frequency of recurrent thromboembolism.
HANS KLAUS BREDDIN, M.D.
Institute of Thrombosis and Vascular Diseases
D-60598 Frankfurt, Germany
breddin@em.uni-frankfurt.de
VIJAY V. KAKKAR, M.D.
Thrombosis Research Institute
London SW3 6LR, United Kingdom
N Engl J Med, Vol. 345, No. 4 July 26, 2001 www.nejm.org 293
The New England Journal of Medicine
Downloaded from nejm.org on September 10, 2017. For personal use only. No other uses without permission.
Copyright 2001 Massachusetts Medical Society. All rights reserved.
Biventricular Pacing in Patients with Heart
Failure
To the Editor: Cardiac output is determined by heart rate
and stroke volume. Since the latter is reduced in patients with
ventricular systolic dysfunction, the cardiovascular reserve
during exercise in patients with heart failure depends heavily
on the increment of the pulse rate. Unfortunately, chrono-
tropic insufficiency is often found in these patients. More-
over, substantial prognostic value has been attributed to this
abnormality.1 We were therefore surprised that Cazeau et
al. (March 2 issue),2 in their report of the Multisite Stimu-
lation in Cardiomyopathies trial, did not provide informa-
tion on the maximal heart rates achieved during exercise be-
fore and after three months of multisite biventricular pacing.
Although this information was not mentioned in the article,
we assume that the minute-ventilation sensor of the im-
planted device was set to provide rate-responsive pacing. If
this was indeed the case, part of the improved exercise ca-
pacity could have resulted from a more physiologic exer-
cise-adapted increase in the heart rate.
VIVIANE CONRAADS, M.D.
CHRISTIAAN VRINTS, M.D., PH.D.
University Hospital Antwerp
2650 Edegem, Belgium
viviane.conraads@uza.uia.ac.be
1. Robbins M, Francis G, Pashkow FJ, et al. Ventilatory and heart rate re-
sponses to exercise: better predictors of heart failure mortality than peak
oxygen consumption. Circulation 1999;100:2411-7.
2. Cazeau S, Leclerq C, Lavergne T, et al. Effects of multisite biventricular
pacing in patients with heart failure and intraventricular conduction delay.
N Engl J Med 2001;344:873-80.
The authors reply:
To the Editor: During the active-pacing period of the six-
month crossover phase, the pulse generator was programmed
in an atrial-synchronous or VDD biventricular pacing mode
with a basic pacing rate of 40 beats per minute (bpm), which
meant that the atrium was used only for sensing and was
never paced. In all patients, the rate-responsive function was
set to off.
We did not observe any significant difference in the max-
imal heart rates at peak exercise at base line (11725 bpm),
at randomization (12524 bpm), at the end of the period
of active pacing (12424 bpm), or at the end of the period
of inactive pacing (12221 bpm).
It clearly appeared that the improvement in exercise tol-
erance with atriobiventricular pacing, as reflected by the in-
creased distance walked in six minutes (P<0.001) and im-
proved peak oxygen uptake (P=0.02), did not reflect the
correction of chronotropic insufficiency but was probably
due to the cardiac-resynchronization effect alone.
In that respect, it is worth recalling the results of our study
of the acute hemodynamic effects of temporary pacing1 in
which we showed that atriobiventricular dual-chamber
(DDD) pacing with an optimized atrioventricular delay re-
sulted in cardiac output that was, on average, 35 percent
greater than that achieved with atrial pacing alone with pre-
served intrinsic conduction. In that study, we used a fixed
pacing rate (10 percent faster than the intrinsic atrial rate) in
patients who, like those included in the more recent study,
 The Ne w E n g l a nd Jo u r n a l o f Me d ic i ne
had severe heart failure and major intraventricular conduc-
tion delay.
JEAN CLAUDE DAUBERT, M.D.
Hpital Pontchaillou
35033 Rennes CEDEX, France
jean-claude.daubert@chu-rennes.fr
SERGE CAZEAU, M.D.
InParis
92210 Saint-Cloud, France
CHRISTOPHE LECLERCQ, M.D.
Hpital Pontchaillou
35033 Rennes CEDEX, France
1. Leclercq C, Cazeau S, Le Breton H, et al. Acute hemodynamic effects
of biventricular DDD pacing in patients with end-stage heart failure. J Am
Coll Cardiol 1998;32:1825-31.
Underuse of Coronary Revascularization
Procedures
To the Editor: Hemingway et al. (March 1 issue)1 con-
clude that, according to criteria set by an expert panel, cor-
onary revascularization procedures are underused; patients
who should have undergone a surgical intervention were
incorrectly treated medically. But what is the basis for the
criteria of the expert panel?
Current recommendations regarding surgery are derived
largely from data from the 1970s and early 1980s,2 which
preceded the development of aggressive medical therapies
both for the management of ischemia (e.g., nitrates and
beta-blockers) and for the reduction of risk factors (especial-
ly cholesterol levels). Even the most recent study comparing
surgical treatment with medical management, conducted in
the mid-1990s, did not include the goal of aggressive lipid
lowering.3 Less than 25 percent of the patients in the med-
ical-treatment group in the study by Hemingway et al. were
receiving hypocholesterolemic agents. Yet it is clear from
other data that aggressive lipid management leads to clear
benefit within six months of initiating therapy.4,5 As com-
pared with coronary angioplasty, aggressive lipid lowering
with 80 mg of atorvastatin (bringing the ratio of total cho-
lesterol to high-density lipoprotein cholesterol down to 2.8)
decreased the rate of ischemic events by 36 percent over a
period of 18 months.4 Since no trial has compared state-
of-the-art medical management (especially aggressive lipid
lowering) with surgery, it is difficult to conclude that cor-
onary revascularization is underutilized, especially for pa-
tients with chronic cardiac symptoms.
GEOFFREY A. MODEST, M.D.
Uphams Corner Health Center
Dorchester, MA 02125
gmodest@partners.org
1. Hemingway H, Crook AM, Feder G, et al. Underuse of coronary re-
vascularization procedures in patients considered appropriate candidates for
revascularization. N Engl J Med 2001;344:645-54.
2. Eagle KA, Guyton RA, Davidoff R, et al. ACC/AHA guidelines for
coronary artery bypass graft surgery: executive summary and recommen-
dations: a report of the American College of Cardiology/American Heart
Association Task Force on Practice Guidelines (Committee to Revise the
294 N Engl J Med, Vol. 345, No. 4 July 26, 2001 www.nejm.org
The New England Journal of Medicine
Downloaded from nejm.org on September 10, 2017. For personal use only. No other uses without permission.
Copyright 2001 Massachusetts Medical Society. All rights reserved.
1991 Guidelines for Coronary Artery Bypass Graft Surgery). Circulation
1999;100:1464-80.
3. Davies RF, Goldberg AD, Forman S, et al. Asymptomatic Cardiac Is-
chemia Pilot (ACIP) study two-year follow-up: outcomes of patients ran-
domized to initial strategies of medical therapy versus revascularization.
Circulation 1997;95:2037-43.
4. Pitt B, Waters D, Brown WV, et al. Aggressive lipid-lowering therapy
compared with angioplasty in stable coronary artery disease. N Engl J Med
1999;341:70-6.
5. The Pravastatin Multinational Study Group for Cardiac Risk Patients.
Effects of pravastatin in patients with serum total cholesterol levels from
5.2 to 7.8 mmol/liter (200 to 300 mg/dl) plus two additional atheroscle-
rotic risk factors. Am J Cardiol 1993;72:1031-7.
To the Editor: Hemingway et al. conclude that patients
who would be considered by an independent panel to be
appropriate candidates for coronary-artery bypass grafting
(CABG) have a greater risk of adverse outcomes if they are
treated medically than if they receive CABG. This conclu-
sion may be valid if the base-line characteristics of the med-
ically and surgically treated patients are uniform. However,
important differences existed in this study.
First, there were significantly more patients with heart
failure and fewer treated with beta-blockers in the medi-
cally treated group. Beta-blockers reduce mortality among
patients with heart failure,1 so the differences in mortality
that were observed in the study could be associated with
their underuse. Second, the medically treated group includ-
ed a higher percentage of patients with diabetes and patients
with previous myocardial infarction. The mortality rate
among patients with diabetes with multivessel coronary
disease is reduced with the use of CABG.2 Previous CABG
also has a cardioprotective effect in patients with diabetes
who have a myocardial infarction.3 Thus, the overall results
may be skewed by a significant effect on a small group of
high-risk patients in whom revascularization has proven
benefit.
KAUSIK K. RAY, M.R.C.P.
PAUL J. SHERIDAN, M.R.C.P.
KOO H. CHAN, M.R.C.P.
University of Sheffield
Sheffield S5 7AU, United Kingdom
k.k.ray@sheffield.ac.uk
1. The Cardiac Insufficiency Bisoprolol Study II (CIBIS-II): a randomised
trial. Lancet 1999;353:9-13.
2. Seven-year outcome in the Bypass Angioplasty Revascularization Inves-
tigation (BARI) by treatment and diabetic status. J Am Coll Cardiol 2000;
35:1122-9.
3. Detre KM, Lombardero MS, Brooks MM, et al. The effect of previous
coronary-artery bypass surgery on the prognosis of patients with diabetes
who have acute myocardial infarction. N Engl J Med 2000;342:989-97.
To the Editor: I am concerned that the article by Hem-
ingway et al. may overstate the benefits of cardiac revascu-
larization because of a possible bias in the data resulting
from the failure to consider the social class of patients. The
group that received medical treatment may have included a
disproportionate number of patients from lower socioeco-
nomic classes. I base this conclusion on the fact that non-
white patients were overrepresented in the medical-treat-
ment groups as compared with the groups assigned to
percutaneous transluminal coronary angioplasty (PTCA) (17
percent vs. 12 percent) and CABG (20 percent vs. 14 per-
cent) and the fact that in the United States nonwhite race
 C ORR ES POND ENCE
is associated with lower social class,1 which I believe is also
the case in the United Kingdom.
There is a well-established association between socioeco-
nomic class and death rates from heart disease.2,3 Including
a disproportionate number of lower-class patients in a study
group can be expected to result in increased death rates,
independently of the treatment the patients receive. Until
the authors can control for differences in social class be-
tween their treatment groups, their conclusions about im-
proved outcomes with revascularization may need to be
muted somewhat.
In addition, it appears that in the United Kingdom, as
in the United States,4 nonwhite patients with heart disease
are referred for revascularization less often than white pa-
tients with a similar level of disease. This finding is as impor-
tant as the difference in treatment outcomes. The adverse
health effects of racial bias in the availability of revascular-
ization procedures far outweigh the effects of underuse of
these procedures among predominantly white, upper-class
patients.
DONALD A. BARR, M.D., PH.D.
Stanford University
Stanford, CA 94305-2160
barr@stanford.edu
1. Bureau of the Census. Money income in the United States: 1999. Cur-
rent population reports. Series P-60. No. 209. Washington, D.C.: Govern-
ment Printing Office, 2000.
2. Black D, Morris JN, Smith C, Townsend P, Whitehead M. Inequalities
in health: the Black Report: the health divide. London: Penguin, 1988.
3. Marmot MG, Kogevinas M, Elston MA. Social/economic status and
disease. Annu Rev Public Health 1987;8:111-35.
4. Peterson ED, Shaw LK, DeLong ER, Pryor DB, Califf RM, Mark DB.
Racial variation in the use of coronary-revascularization procedures: are the
differences real? Do they matter? N Engl J Med 1997;336:480-6.
To the Editor: Hemingway et al. report data suggesting
that the use of explicit measures of appropriateness may
result in a more judicious use of revascularization proce-
dures. We wonder whether patients in this study were as-
signed to the CABG group on an intention-to-treat basis
or on the basis of the treatment they received. Specifically,
if a patient who was going to receive CABG died before he
was able to undergo surgery, in which group was he count-
ed? We note in Figure 1 of the article that large numbers of
deaths or nonfatal myocardial infarctions occurred soon after
the initial angiography was performed, particularly in the
medically treated group. Thus, the failure to categorize the
patients according to the intention-to-treat principle would
bias the results of this study against medical treatment.
AARIF Y. KHAKOO, M.D.
DARIUS A. RASTEGAR, M.D.
Johns Hopkins Bayview Medical Center
Baltimore, MD 21224
akhakoo@jhmi.edu
The authors reply:
To the Editor: The correspondents raise important issues
in interpreting the finding in the Appropriateness of Cor-
onary Revascularization study of the underuse of CABG
after angiography. However, these further considerations
N Engl J Med, Vol. 345, No. 4 July 26, 2001 www.nejm.org 295
The New England Journal of Medicine
Downloaded from nejm.org on September 10, 2017. For personal use only. No other uses without permission.
Copyright 2001 Massachusetts Medical Society. All rights reserved.
(the subjects of detailed reports published elsewhere) do
not alter the main conclusion of the study. We report here
hazard ratios for death among patients in whom CABG was
deemed appropriate, comparing patients who received no
revascularization with those who underwent CABG. Haz-
ard ratios greater than 1.0 denote an underuse of CABG.
In order to address the suggestion made by both Mod-
est and Ray et al. that optimizing medical management
might alter our findings, we carried out subgroup analyses
among patients who were taking a statin or a beta-blocker
and calculated hazard ratios of 4.47 (P=0.002) and 3.79
(P<0.001), respectively. The similarity between these haz-
ard ratios and the ratio of 4.96 we reported suggests that
optimizing medical management may make little difference
in the effect of the underuse of CABG on mortality.
Ray and colleagues note an excess of diabetes, myocar-
dial infarction, and heart failure among the medically treat-
ed group. Adjustment for these factors (in Tables 3, 4, and
5 of our article) had little effect on our results; analyses that
excluded patients with any of these coexisting conditions
actually found a stronger effect of the underuse of CABG
on mortality (hazard ratio, 5.97; P<0.001).
As Barr suggested, we have now controlled for social class
and race; neither of these factors attenuated the effect of the
underuse of CABG on mortality, with hazard ratios of 4.49
(P<0.001) and 3.89 (P<0.001), respectively. We found no
evidence that South Asian patients were less likely than
whites either to be deemed appropriate candidates for re-
vascularization or to be referred for a revascularization pro-
cedure.
The primary, prestated intention of our study was to
compare the clinical outcomes among patients who received
a given treatment for coronary disease with the outcomes
among those who did not, according to prespecified levels
of appropriateness. Because the timeliness with which pa-
tients undergo a revascularization procedure is a crucial con-
sideration for any investigation of the magnitude of under-
use, any patient who died before receiving revascularization
was included in the analyses as a member of the medical
group. To do otherwise would be to create a conservative
bias. However, our results are robust enough to support an
intention-to-treat analysis of the type suggested by Khakoo
and Rastegar. Among those classified as appropriate candi-
dates for CABG, the patients for whom medical treatment
was intended had higher mortality rates than those for whom
CABG was intended, regardless of the actual therapy they
received (hazard ratio, 2.05; P=0.004).
The challenge now is to determine the extent to which
clinical outcomes are improved when suitably updated ap-
propriateness criteria for angiography1 and revascularization
are used to support clinical decisions in routine practice; the
application of these criteria should not be delayed.2
HARRY HEMINGWAY, M.R.C.P.
ANGELA M. CROOK, M.SC.
Kensington & Chelsea and Westminster Health Authority
London W2 6LX, United Kingdom
harry.hemingway@ha.kcw-ha.nthames.nhs.uk
ADAM D. TIMMIS, M.D., F.R.C.P.
Barts and the London NHS Trust
London E2 9JX, United Kingdom
1. Hemingway H, Crook AM, Banerjee S, et al. Hypothetical ratings of
coronary angiography appropriateness: are they associated with actual an-
 The Ne w E n g l a nd Jo u r n a l o f Me d ic i ne
giographic findings, mortality, and revascularisation rate? The ACRE study.
Heart 2001;85:672-9.
2. Shekelle PG. Are appropriateness criteria ready for use in clinical prac-
tice? N Engl J Med 2001;344:677-8.
Cutaneous Squamous-Cell Carcinoma
To the Editor: In their excellent review article on cutane-
ous squamous-cell cancers (March 29 issue),1 Alam and Rat-
ner mention sunscreen use but not the prevention of these
cancers in certain populations of patients at high risk.
Among recipients of solid-organ transplants, the incidence
of cutaneous squamous-cell carcinoma may be more than
100 times the incidence in the general population.2 More-
over, as many as 6 percent of renal-allograft recipients with
skin cancer may die of metastases.
Prophylactic measures, including the daily use of topical
tretinoin, enhance the number and function of dendritic
cells within the skin, a change seen in association with a
decrease in the formation of new skin cancers.3,4 In selected
patients, the addition of oral retinoids may provide further
benefit.4 However, we have observed acute allograft rejec-
tion during the use of high doses of oral retinoids, a prob-
lem that may be due to the ability of these agents to induce
interferon-g production,5 which may play a part in abrogat-
ing allograft tolerance. We advocate the use of 13-cis-reti-
noic acid at a dose of 10 mg every day or every other day as
an adjunct to the use of topical tretinoin; adverse effects have
not been observed at this dose. In the future, novel topical
agents with potent local immune-enhancing effects, such
as resiquimod, may prove to be of even greater benefit.
ALAIN H. ROOK, M.D.
MICHAEL SHAPIRO, M.D.
University of Pennsylvania School of Medicine
Philadelphia, PA 19104
arook@mail.med.upenn.edu
1. Alam M, Ratner D. Cutaneous squamous-cell carcinoma. N Engl J Med
2001;344:975-83.
2. Penn I. Cancer is a long-term hazard of immunosuppressive therapy.
J Autoimmun 1988;1:545-58.
3. Chen G, Kang K, Kang S, et al. Differential induction of IL-12 p40
and IL-10 mRNA in human Langerhans cells and keratinocytes by in vivo
occlusion, vehicle, and all-trans retinoic acid. J Cutan Med Surg 1996;1:
74-80.
4. Rook AH, Jaworsky C, Nguyen T, et al. Beneficial effect of low-dose
systemic retinoid in combination with topical tretinoin for the treatment
and prophylaxis of premalignant and malignant skin lesions in renal trans-
plant recipients. Transplantation 1995;59:714-9.
5. Fox FE, Kubin M, Cassin M, et al. Retinoids synergize with interleu-
kin-2 to augment IFN-g and interleukin-12 production by human periph-
eral blood mononuclear cells. J Interferon Cytokine Res 1999;19:407-
15.
To the Editor: With regard to the article by Alam and
Ratner on cutaneous squamous-cell carcinoma, I would
like to draw attention to the depiction of skin carcinogen-
esis in Figure 1. Actinic keratoses do not begin in the up-
per layers of the epidermis, but within the basal layer
that is, within the only layer where cells replicate. Suprabasal
keratinocytes cannot replicate. Keratinocytes acquire ultra-
violet-induced p53 mutations while they are within the
basal layer. In conventional immunohistochemical sections,
296 N Engl J Med, Vol. 345, No. 4 July 26, 2001 www.nejm.org
The New England Journal of Medicine
Downloaded from nejm.org on September 10, 2017. For personal use only. No other uses without permission.
Copyright 2001 Massachusetts Medical Society. All rights reserved.
they can be seen in the upper layers as a consequence of up-
ward migration a feature common to all keratinocytes.
LORENZO CERRONI, M.D.
University of Graz
A-8036 Graz, Austria
lorenzo.cerroni@kfunigraz.ac.at
To the Editor: As a radiation oncologist, I generally see
patients with skin cancer of the head and neck, where sur-
gery can be disfiguring. The majority of patients are eld-
erly, and thus do not face the theoretical risk (probably no
greater than 1 to 2 percent over a period of 30 years) of a
secondary cancer. However, the review article does little to
reassure practitioners that radiotherapy is a reasonable and
very effective alternative to surgery. A previous review article1
stated that radiation therapy maximizes tissue preserva-
tion, that it is especially advantageous in elderly, debili-
tated, or other patients at higher risk of surgical complica-
tions, and that it is most appropriate for lesions at sites
where tissue preservation is essential. Another review arti-
cle on curative radiotherapy for skin cancers reported a 95
percent rate of local control for eyelid tumors and a 93 per-
cent rate of local control for tumors of the nose.2 Other
studies have reported excellent local control, in excess of
90 percent, with radiotherapy for squamous-cell tumors of
the pinna and medial fleshy canthus.3
JONDAVID POLLOCK, M.D., PH.D.
Schiffler Cancer Center
Wheeling, WV 26003
jpollock@wheelinghosp.com
1. Preston DS, Stern RS. Nonmelanoma cancers of the skin. N Engl J Med
1992;327:1649-62.
2. Morrison WH, Garden AS, Ang KK. Radiation therapy for nonmela-
noma skin carcinomas. Clin Plast Surg 1997;24:719-29.
3. Petrovich Z, Kuisk H, Langholz B, et al. Treatment results and patterns
of failure in 646 patients with carcinoma of the eyelids, pinna, and nose.
Am J Surg 1987;154:447-50.
The authors reply:
To the Editor: We agree with Rook and Shapiro that the
prevention of squamous-cell carcinoma in patients who have
received a solid-organ transplant is an essential part of man-
agement. Topical and systemic retinoids are certain to have
an increasing role in tumor prophylaxis in these patients, and
we look forward to the development of new topical immu-
nomodulatory agents to treat patients with human papillo-
mavirus-associated and nonhuman papillomavirus-associ-
ated squamous-cell carcinoma.
Cerroni is correct in stating that actinic keratoses begin
in the basal layer of the epidermis as a result of p53 muta-
tions induced by ultraviolet radiation. We tried to simplify
this concept in our diagram, but unfortunately, the cells
with dysfunctional p53 genes are shown too high in the
epidermis. The abnormal cell population should instead be
shown initially in the basal layer. After ultraviolet radiation,
these cells undergo clonal expansion into the upper por-
tions of the epidermis, instead of exhibiting downward
growth.
We disagree with Pollocks assertion that our article
 C ORR ES POND ENCE

does little to reassure practitioners that radiotherapy is a
reasonable and very effective alternative to surgery. In fact,
we state that radiation as a primary treatment may pro-
vide favorable functional and cosmetic results for proper-
ly selected patients with squamous-cell carcinoma and that
radiation may be used in combination with other types of
therapy to treat aggressive or recurrent lesions. It is impor-
tant to note that nonmelanoma skin cancer, which is dis-
cussed in a previous review article, consists principally of
basal-cell carcinomas.1 Although radiation therapy may be
useful to treat primary basal-cell carcinomas in elderly or
debilitated patients or in other patients at increased risk for
surgical complications, basal-cell carcinomas and squamous-
cell carcinomas have far different rates of metastasis. Al-
though both populations of tumors may become more ag-
gressive if they recur after radiation therapy, recurrent squa-
mous-cell carcinomas carry a much greater risk of metastasis
than recurrent basal-cell carcinomas and are also associat-
ed with a higher mortality rate.2
DSIRE RATNER, M.D.
MURAD ALAM, M.D.
Columbia Presbyterian Medical Center
New York, NY 10032
dr221@columbia.edu
id. Biopsy specimens of the peritoneum revealed lympho-
histiocytic nodules and fibrosis (Fig. 1), but no polarizing
material was found and no microorganisms were identified
by auramine and Gomoris methenamine silver staining. De-
spite therapy with prednisone (60 mg daily for four weeks),
the patients ascites recurred. After she stopped taking the
drug, the peritoneal catheter was removed and a ventricu-
loatrial shunt was placed. Within two weeks, her abdomi-
nal distention disappeared. No ascites was evident on ul-
trasonography one and a half years later.
Impaired absorption of cerebrospinal fluid across an in-
flamed peritoneum has been proposed as the cause of cere-
brospinal fluid ascites. Diversion of cerebrospinal fluid from
the abdomen is an effective treatment.1 A patient from Spain
with pathological features similar to those in our patient
has been reported.2 Silicone and its constituents do not cause
specific immune responses,3 but silicone shunts can degrade
over time and elicit nonspecific tissue inflammation.4
GEORGE F. LONGSTRETH, M.D.
NAOMI R. BUCKWALTER, M.D.
Kaiser Permanente Medical Care Program
San Diego, CA 92120
george.f.longstreth@kp.org

1. Preston DS, Stern RS. Nonmelanoma cancers of the skin. N Engl J Med
1992;327:1649-62.
2. Johnson TM, Rowe DE, Nelson BR, Swanson NA. Squamous cell car-
cinoma of the skin (excluding lip and oral mucosa). J Am Acad Dermatol
1992;26:467-84.

Sterile Cerebrospinal Fluid Ascites and Chronic

Peritonitis

To the Editor: Sterile ascites after ventriculoperitoneal

shunting is rare and usually of unknown cause.1 We describe
a patient in whom an inflammatory reaction to silicone
tubing used in the creation of a shunt was a possible cause
of sterile ascites. N
A 28-year-old woman presented with abdominal disten-
tion of three months duration. She had undergone ven-
triculoperitoneal shunting at six months of age for hydro-
cephalus and subsequently required multiple revisions and
replacements of siliconeelastomer shunts. She had no his-
tory of shunt infection or of abdominal or pelvic infection
or surgery. She took valproic acid and carbamazepine for
seizures. The results of routine blood and urine tests were
normal. Computed tomography of the abdomen and pel-
vis showed extensive ascites. Paracentesis revealed clear flu-
id with 150 leukocytes per cubic millimeter (95 percent
mononuclear cells) and a protein level of 2.8 g per deciliter;
cytologic examination showed no malignant cells. The se-
rumascites albumin gradient was 2.5 g per deciliter. No
organisms grew in cultures of ascitic fluid. The results of ab- F
dominal ultrasonography, echocardiography, and a liver bi-
opsy were normal.
Despite diuretic therapy, the patient required multiple
therapeutic paracenteses, with a total of 18 liters drained
over a period of six months. Laparoscopy revealed exten-
sive abdominopelvic adhesions, numerous tiny nodules on
the visceral and parietal peritoneum and inferior surface of Figure 1. Biopsy Specimen of the Parietal Peritoneum, Showing
the liver, and a fragment of a shunt tube. The laparoscopist a Lymphohistiocytic Nodule (N) and Fibrosis (F) (Hematoxylin
extracted the fragment and removed 5 liters of ascitic flu- and Eosin, 100).

N Engl J Med, Vol. 345, No. 4 July 26, 2001 www.nejm.org 297

The New England Journal of Medicine

Downloaded from nejm.org on September 10, 2017. For personal use only. No other uses without permission.
Copyright 2001 Massachusetts Medical Society. All rights reserved.
 The Ne w E n g l a nd Jo u r n a l o f Me d ic i ne

1. Yukinaka M, Nomura M, Mitani T, et al. Cerebrospinal ascites devel-
oped 3 years after ventriculoperitoneal shunting in a hydrocephalic patient.
Intern Med 1998;37:638-41.
2. Prez Pea F, Aparicio Campillo G, Lpez Asenjo JA, et al. Ascitis
por acmulo de lquido cefalorraqudeo. Rev Clin Esp 1990;187:128-
30.
3. Saxon A. The antigen that wasnt silicone. Clin Immunol 2000;95:
171-2.
4. Del Bigio MR. Biological reactions to cerebrospinal fluid shunt devices:
a review of the cellular pathology. Neurosurgery 1998;42:319-26.
Correspondence Copyright 2001 Massachusetts Medical Society.

298 N Engl J Med, Vol. 345, No. 4 July 26, 2001 www.nejm.org

The New England Journal of Medicine

Downloaded from nejm.org on September 10, 2017. For personal use only. No other uses without permission.
Copyright 2001 Massachusetts Medical Society. All rights reserved.