Oral Clonidine in Postmenopausal Patients with Breast Cancer

Experiencing Tamoxifen-Induced Hot Flashes: A University of

Rochester Cancer Center Community Clinical Oncology
Program Study
Kishan J. Pandya, MD; Richard F. Raubertas, PhD; Patrick J. Flynn, MD; Harry E. Hynes, MD;
Richard J. Rosenbluth, MD; Jeffrey J. Kirshner, MD; H. Irving Pierce, MD; Vladimir Dragalin, PhD;
and Gary R. Morrow, PhD, MS

Background: Hot flashes are the most frequently re-

ported side effect of tamoxifen treatment. Although hor-
mones are an effective treatment, their safety is question-
H ot flashes—a collective term that includes
many vasomotor symptoms, such as a feeling
of warmth, redness of the face and upper body,
able in women with breast cancer. It is therefore important
sweating, and dizziness—are the most frequent meno-
to evaluate nonhormonal treatments for hot flashes.
pausal symptom. Approximately 60% to 70% of
Objective: To evaluate the effectiveness of oral clonidine women undergoing the climacteric report hot flashes
for control of hot flashes associated with tamoxifen ther-
(1). Hot flashes are also the most common side effect
apy in postmenopausal women with breast cancer.
associated with use of the antiestrogen tamoxifen; in
Design: Randomized, double-blind, placebo-controlled a recent study, 57% of 1318 women with breast
clinical trial.
cancer who were receiving the drug reported hot
Setting: University of Rochester Cancer Center Commu- flashes (2). Although most women tolerate this symp-
nity Clinical Oncology Program. tom and are able to remain active, others find it
Patients: 194 postmenopausal women with breast cancer distressing. Hot flashes have been reported to inter-
who were receiving adjuvant tamoxifen therapy. fere with daily activities and sleep, and some women
Intervention: Oral clonidine hydrochloride, 0.1 mg/d, or have considered discontinuing therapy or have been
placebo for 8 weeks. less compliant with their daily tamoxifen regimen.
Measurements: In a daily diary, patients recorded num- These problems have become more pronounced be-
ber, duration, and severity of hot flashes and overall cause long-term tamoxifen therapy is now advocated
quality-of-life score (on a 10-point scale) during a 1-week for early-stage breast cancer. It is important to ad-
baseline period and during the 4th, 8th, and 12th weeks dress the continuing problem of vasomotor side ef-
of the study. fects associated with tamoxifen therapy.
Results: Patients in the placebo and treatment groups The pathophysiology underlying hot flashes is not
were similar in age, duration of tamoxifen use, reported entirely clear. Casper and Yen (3) have hypothe-
frequency and duration of hot flashes at baseline, and sized that physiologic levels of estrogen and proges-
dropout rates. One hundred forty-nine patients completed terone maintain endogenous opioid peptide concen-
12 weeks of follow-up. The mean decrease in hot flash
trations in the hypothalamus and brain stem. At
frequency was greater in the clonidine group than in the
placebo group after 4 weeks of treatment (37% compared
menopause, these concentrations decrease with de-
with 20% [95% CI for difference, 7% to 27%]) and 8 creasing estrogen levels, which releases nonadrener-
weeks of treatment (38% compared with 24% [CI for gic activity from its usual tonic inhibition and causes
difference, 3% to 27%]). Patients receiving clonidine were nonadrenergic hyperactivity. This leads to inappro-
more likely than patients receiving placebo to report dif- priate activation of the heat loss mechanism in the
ficulty sleeping (41% compared with 21%; P ⫽ 0.02). A medial preoptic area and subsequent hot flashes.
significant difference was seen in the mean change in Clonidine hydrochloride is a centrally active
quality-of-life scores (0.3 points in the clonidine group ␣-adrenergic agonist that reduces vascular reactivity
compared with ⫺0.2 points in the placebo group; P ⫽ 0.02)
and is currently used to treat hypertension. Its re-
at 8 weeks, although the median difference was 0 in both
groups.
ported side effects include dry mouth and sleepi-
ness. Several studies have examined its effectiveness
Conclusion: Oral clonidine, 0.1 mg/d, is effective against
in controlling hot flashes caused by natural or sur-
tamoxifen-induced hot flashes in postmenopausal women
with breast cancer.
gical menopause. Wren and Brown (4) reported no
benefit of clonidine therapy in a placebo-controlled
trial, but Clayden and colleagues (5) and Edington
Ann Intern Med. 2000;132:788-793. and coworkers (6) demonstrated significant benefit
For author affiliations, current addresses, and contributions, see in patients receiving oral clonidine. In a dose-esca-
end of text. lation study by Laufer and coworkers (7), 10 women
788 © 2000 American College of Physicians–American Society of Internal Medicine

Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19956/ by a University of California San Diego User on 01/16/2017

 received two daily doses of clonidine, starting with cer Prevention and Control, Bethesda, Maryland,
0.1 mg/d and increasing to 0.2 mg/d and 0.4 mg/d. funded all aspects of the study as part of their
Each dose level was maintained for 2 weeks. Four CCOP. The study was designed, conducted, peer
of the 10 participants withdrew because of side ef- reviewed, and approved at the University of Roch-
fects: 1 at 0.1 mg/d, 1 at 0.2 mg/d, and 2 at 0.4 ester Cancer Center, a designated research base for
mg/d. Increasing benefit was noted with increasing CCOP. The University of Rochester Institutional
dose: At the maximum dose, hot flashes decreased Review Board approved the study.
by 46%. In a more recent study, transdermal
clonidine was found to be of some benefit for ta- Treatment
moxifen-induced hot flashes (8). Patients were randomly assigned to receive oral
A double-blind crossover pilot study of clonidine, clonidine, 0.1 mg, or placebo at bedtime for 8
0.1 mg, or placebo given at bedtime was conducted weeks. Randomization was done by using a com-
at the University of Rochester Cancer Center in puter program maintained by the University of
Rochester, New York (9). Fifteen women receiving Rochester Department of Biostatistics and was
tamoxifen for breast cancer entered the study. Of stratified by time since menopause (ⱕ3 years, ⬎3
the 13 evaluable women, 9 (69%) preferred clonidine, years), duration of tamoxifen therapy (ⱕ1 year, ⬎1
2 (15%) had no preference, and 2 (15%) preferred year), and baseline frequency of hot flashes (⬍10
placebo. No adverse side effects, including any related per day, ⱖ10 per day). Doses were not modified
to blood pressure, were reported. The clonidine and during the study. Patients or physicians could dis-
placebo groups did not differ significantly with re- continue treatment for any reason. If treatment was
gard to number of hot flashes. However, because discontinued, the reason was recorded on the appro-
the sample was too small and the duration of treat- priate study forms.
ment too short, differences between treatments
could not be reliably detected. We performed a Treatment Evaluation
larger randomized, controlled trial that studied the Self-report methods were used because the pa-
effect of clonidine on hot flashes in women receiv- tient’s evaluation of symptoms was of principal im-
ing tamoxifen for breast cancer. portance and greatest clinical relevance (10). We
obtained data on hot flashes by asking patients to
Methods keep a daily diary for 1 week at baseline, during the
4th and 8th weeks of treatment, and during the 4th
Our study was conducted by the University of week after the end of treatment (the 12th week
Rochester Cancer Center and its affiliates under the after randomization). The diary consisted of self-
auspices of the National Cancer Institute’s Commu- administered questionnaires on the number, dura-
nity Clinical Oncology Program (CCOP). Patients tion, and severity of hot flashes. Severity was re-
who met eligibility criteria were recruited from clin- corded as mild, moderate, severe, or very severe,
ical practices of medical oncologists from participat- similar to that reported by the North Central Can-
ing CCOPs. Postmenopausal women who had been cer Treatment Group (8). A separate symptom
receiving adjuvant tamoxifen therapy for breast can- checklist was used to monitor 18 potential side ef-
cer for at least 1 month and reported at least one fects, as adapted from the drug surveillance litera-
hot flash per day were eligible. Premenopausal ture and used in previous studies (11). Potential
women and women receiving concurrent chemother- symptoms were difficulty sleeping, decreased appe-
apy or other endocrine therapy for breast cancer tite, constipation, diarrhea, hair loss, headache, eye
were ineligible. Patients receiving hypertension ther- sensitivity to light, dry mouth, night sweats, pain,
apy or concurrent treatment with monoamine oxi- nausea, vomiting, fatigue, stress incontinence,
dase inhibitors; L-dopa; piribedil; tricyclic antide- change in sense of taste or smell, drowsiness, dizzi-
pressants; or sedatives, such as benzodiazepines or ness, and hot flashes. We also asked patients to rate
barbiturates, were also ineligible. Patients with cor- their lives on a scale from 1 (worst possible life) to
onary insufficiency, recent history of myocardial in- 10 (best possible life) at each assessment. A brief
farction (within the past 3 months), symptomatic history and physical examination, including blood
cardiac disease, peripheral or cerebrovascular dis- pressure measurement, were done at each assessment.
ease, syncope, or symptomatic hypotension were ex-
cluded, as were those with a history of allergy or Statistical Analysis
adverse reaction to clonidine. Normal hepatic and Four measures of hot flash symptoms were used
renal function were required. Each patient signed to assess treatment effectiveness: frequency, mean
an informed consent document approved by the severity grade (calculated by assigning scores of 1, 2,
institutional review board of the participating affiliate. 3, and 4, respectively, to mild, moderate, severe,
The National Cancer Institute, Division of Can- and very severe hot flashes), a combined score ob-
16 May 2000 • Annals of Internal Medicine • Volume 132 • Number 10 789

Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19956/ by a University of California San Diego User on 01/16/2017

 minus its baseline severity. Between-group differ-
ences were compared by using the Wilcoxon test.

Results

The flow of patients is shown in Figure 1. We

randomly assigned 198 women to receive clonidine
or placebo. Four women did not provide baseline
information about hot flashes and thus were consid-
ered unevaluable. The 194 evaluable patients in the
clonidine and placebo groups were similar with re-
spect to mean age (53 years compared with 55 years
[range, 35 to 77 years]; P ⫽ 0.15), duration of ta-
moxifen use (⬎1 year, 39% compared with 35%;
P ⬎ 0.2), and time since menopause (⬎3 years, 58%
compared with 62%; P ⬎ 0.2). Forty-five of 194 pa-
tients (23%) did not complete the full 12 weeks of
the study, and 31 (16%) did not complete the
8-week intervention period. Dropout rates were
similar in the two groups and were due to patient
preference. Baseline characteristics (age, hot flash
symptoms, duration of tamoxifen therapy, time since
menopause, and quality-of-life score) were generally
similar in patients who completed the study and
those who did not. Only one difference was statis-
tically significant: Patients who dropped out had a
shorter mean duration of hot flashes at baseline
than did the 149 patients who completed the study.
During the baseline week, the mean frequency of
hot flashes among the two groups was 8.0 (median,
Figure 1. Flow of participants. *Two patients were ineligible, and 6.7) per day in the clonidine group and 7.4 (median,
three declined to participate.
6.3) per day in the placebo group. Mean severity
grades were 2.2 and 2.1, respectively (median, 2.1)
tained by multiplying the mean frequency by the (Table). The median duration of hot flashes was
mean severity grade, and mean duration. Symptoms approximately 3 minutes in both groups. The mean
during the 4th, 8th, and 12th weeks were expressed duration in the clonidine group was distorted by a
as percentage changes from the baseline week for single patient who reported very long durations
each woman, and the Wilcoxon test was used to test (mean, ⬎88 minutes). Exclusion of this patient
for differences between the clonidine and placebo caused only minor changes in the differences be-
groups. Patients’ ratings of quality of life at each tween treatment groups. Mean quality-of-life scores
assessment were expressed as changes from the were between 7.0 and 8.0 for both groups. No sta-
baseline score, and the Wilcoxon test was used to tistically significant differences were observed be-
test for a difference between the clonidine and pla- tween treatment groups at baseline.
cebo groups. Repeated measures of analysis of vari- The Table and Figure 2 show changes in the
ance were used to test for differences in treatment outcome measures from baseline by treatment
effects between weeks 4 and 8. Two-way analysis of group and week. Both groups reported milder hot
variance was used to test for differences in treat- flash symptoms during treatment than during the
ment effect between subgroups defined by time baseline period, but the reduction was greater in the
since menopause, duration of tamoxifen therapy, or clonidine group. Patients receiving clonidine had
baseline frequency of hot flashes. We used SAS (SAS significantly greater benefit with respect to fre-
Institute, Inc., Cary, North Carolina) and S-PLUS quency and severity of hot flashes at weeks 4 and 8;
(MathSoft, Inc., Seattle, Washington) statistical soft- 95% CIs for the between-group difference in fre-
ware. All P values are two-sided. quency changes were 7% to 27% at week 4 and 3%
Patients reported occurrence and severity of 18 to 27% at week 8. Although a nonsignificant differ-
side effects on a scale from 0 (not at all severe) to ence between the clonidine and placebo groups was
4 (extremely severe). The data were summarized as observed at week 12 (4 weeks after the end of treat-
the peak severity of each symptom during treatment ment), the percentage change was greater in the
790 16 May 2000 • Annals of Internal Medicine • Volume 132 • Number 10

Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19956/ by a University of California San Diego User on 01/16/2017

 Table. Comparison of Clonidine and Placebo Groups by Week of Follow-up*

Outcome Clonidine Group Placebo Group P

Value†
Patients Mean ⫾ SE Median Patients Mean ⫾ SE Median
n n

Hot flashes per day

Baseline, n 97 8.0 ⫾ 0.6 6.7 97 7.4 ⫾ 0.5 6.3 ⬎0.2
Percentage change at week 4 90 ⫺37.0 ⫾ 3.5 ⫺34.2 91 ⫺20.1 ⫾ 3.8 ⫺21.1 0.001
Percentage change at week 8 84 ⫺38.4 ⫾ 4.2 ⫺36.7 79 ⫺23.6 ⫾ 4.3 ⫺16.7 0.006
Percentage change at week 12 73 ⫺23.8 ⫾ 5.2 ⫺25.0 76 ⫺13.9 ⫾ 3.9 ⫺15.0 0.09
Hot flash severity‡
Baseline 97 2.2 ⫾ 0.1 2.1 97 2.1 ⫾ 0.1 2.1 ⬎0.2
Percentage change at week 4 90 ⫺14.9 ⫾ 2.9 ⫺11.7 91 ⫺9.7 ⫾ 2.6 ⫺8.5 0.18
Percentage change at week 8 84 ⫺21.0 ⫾ 3.3 ⫺17.3 79 ⫺13.1 ⫾ 3.4 ⫺10.5 0.08
Percentage change at week 12 73 ⫺14.6 ⫾ 3.8 ⫺9.3 76 ⫺6.1 ⫾ 3.3 ⫺8.3 ⬎0.2
Hot flash score§
Baseline 97 18.3 ⫾ 1.6 13.3 97 16.6 ⫾ 1.2 12.6 ⬎0.2
Percentage change at week 4 90 ⫺42.2 ⫾ 4.4 ⫺42.6 91 ⫺23.7 ⫾ 4.6 ⫺24.1 0.002
Percentage change at week 8 84 ⫺45.0 ⫾ 4.9 ⫺50.2 79 ⫺26.4 ⫾ 5.5 ⫺25.8 0.006
Percentage change at week 12 73 ⫺28.6 ⫾ 5.8 ⫺26.6 76 ⫺16.1 ⫾ 4.9 ⫺17.5 0.07
Hot flash duration
Baseline, min 97 8.4 ⫾ 4.1 2.7 97 4.1 ⫾ 0.3 3.1 ⬎0.2
Percentage change at week 4 90 ⫺15.5 ⫾ 4.2 ⫺11.3 89 ⫺2.8 ⫾ 5.8 ⫺1.5 0.11
Percentage change at week 8 84 ⫺23.2 ⫾ 4.3 ⫺16.7 78 ⫺9.2 ⫾ 5.7 ⫺16.8 0.18
Percentage change at week 12 73 ⫺21.5 ⫾ 5.0 ⫺23.2 75 17.0 ⫾ 19.5 ⫺4.4 0.023
Quality-of-life score㛳
Baseline 96 7.2 ⫾ 0.2 7.0 97 7.7 ⫾ 0.2 8.0 0.08
Change at week 4 90 0.4 ⫾ 0.1 0.0 88 ⫺0.3 ⫾ 0.2 0.0 0.003
Change at week 8 83 0.3 ⫾ 0.2 0.0 77 ⫺0.2 ⫾ 0.2 0.0 0.022
Change at week 12 75 0.3 ⫾ 0.2 0.0 78 ⫺0.1 ⫾ 0.2 0.0 ⬎0.2

* All changes are relative to baseline.

† Two-sided Wilcoxon test comparing clonidine and placebo groups.
‡ On a scale of 1 to 4.
§ Obtained by multiplying the mean frequency by the mean severity grade.
㛳 On a scale of 1 to 10.

clonidine group. A significant difference was seen in 0.022), and 2) the last-observation-carried-forward
the mean changes in quality-of-life score at 8 weeks analysis identified additional significant differences
(0.3 points in the clonidine group compared with that showed greater benefit with clonidine for hot
⫺0.2 points in the placebo group), although the flash frequency, hot flash score, and quality-of-life
median difference was 0 in both groups. Treatment score at 12 weeks. Neither of the alternate analyses
effects did not differ significantly between the 4- and is preferable to the available-data analysis, but both
8-week follow-ups for any of the outcome measures illustrate that the main conclusions of the study are
(P ⬎ 0.2 for all treatment-by-week interactions). not sensitive to the handling of missing data, espe-
To analyze treatment effects, we used all avail- cially during the 8-week intervention period.
able observations at each follow-up. To assess the We also tested whether the effects of clonidine at
sensitivity of the results to alternate ways of han- weeks 4 and 8 differed by the stratification variables,
dling missing values, we repeated the primary anal- which interacted with treatment effect in two statis-
ysis with two variations. In the first, we excluded the
tically significant ways. At week 4, the effect of
45 patients who dropped out before the end of the
clonidine on hot flash duration was significantly
12-week study; that is, only data from those who
greater for women receiving tamoxifen for 1 year
completed the study were analyzed. In the second,
than for those receiving it for more than 1 year
missing values at weeks 8 or 12 were replaced by
(P ⫽ 0.024). At week 8, the clonidine effect on hot
the preceding nonmissing value for the patient (last
observation carried forward). Because outcomes flash score was significantly greater for women 3
were defined as changes from baseline, we had no years past menopause than for women more than 3
value to carry forward for missing data at week 4. years past menopause (P ⫽ 0.048). These results
However, dropout rates at week 4 were less than should be interpreted with caution because we tested
10% and were nearly identical in the two treatment 30 potential interactions and thus could expect to ob-
groups. For both alternate analyses, the patterns serve 1 or 2 nominally significant effects by chance.
and magnitudes of treatment effects were similar to Patients self-reported the occurrence and severity
those for the reported analyses. Only the following of potential side effects. Difficulty sleeping was the
changes in statistical significance were seen: 1) The only variable that differed significantly between
analysis of persons who completed the study showed groups; 41% of those in the clonidine group and
less evidence of a treatment effect on quality-of-life 21% of those in the placebo group reported in-
score at 8 weeks (P ⫽ 0.082 compared with P ⫽ creased difficulty (P ⫽ 0.02). Blood pressure was
16 May 2000 • Annals of Internal Medicine • Volume 132 • Number 10 791

Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19956/ by a University of California San Diego User on 01/16/2017

 monitored at each assessment visit, and no adverse for hot flashes in survivors of breast cancer (16). A
effect was noted. statistically significant reduction in hot flash fre-
quency was seen, but the clinical magnitude was
considered marginal because patients experienced
Discussion
only one fewer hot flash per day.
Hot flashes are a bothersome side effect of ta- Our trial demonstrates that oral clonidine can
moxifen therapy in postmenopausal women with decrease the frequency of tamoxifen-induced hot
breast cancer. For patients with breast cancer, phy- flashes for at least 8 weeks. This is consistent with
sicians are usually reluctant to prescribe estrogens the reported effectiveness of clonidine in women
because of the possibility of detrimental effects. Low who experienced natural menopause (5, 6). Trans-
doses of progestational agents, such as megestrol, dermal clonidine has also been found to reduce the
have been reported to control hot flashes (12–15). It frequency and, to a lesser extent, the severity of
is uncertain, however, whether these agents are safe tamoxifen-induced hot flashes over a 4-week period
in women receiving tamoxifen, because both classes (8). Our results are similar to those seen with trans-
of drugs are potentially thrombogenic. Therefore, dermal clonidine over the first 4 weeks of adminis-
effective nonhormonal treatments are needed for tration. In addition, effectiveness was preserved over
tamoxifen-induced hot flashes. the 8-week duration of our study; a median of ap-
A recent prospective study evaluated vitamin E proximately 2.2 fewer hot flashes per day occurred

Figure 2. Changes in hot flash symptoms and quality-of-life score compared with baseline values, by treatment group and study week. For
each week, the left-hand boxplot with squares represents the clonidine group and the right-hand boxplot with circles represents the placebo group. One
observation at week 12 in the placebo group is not included in the plot of hot flash duration because the patient reported a mean duration of 1 minute at
baseline and 15 minutes at 12 weeks (an increase of 1400%).

792 16 May 2000 • Annals of Internal Medicine • Volume 132 • Number 10

Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19956/ by a University of California San Diego User on 01/16/2017

 in the treatment group compared with 1.2 fewer hot Requests To Purchase Bulk Reprints (minimum, 100 copies): Bar-
bara Hudson, Reprints Coordinator; phone, 215-351-2657; e-mail,
flashes per day in the placebo group. Clonidine had [email protected].
a small beneficial effect on hot flash duration and
severity. Mean quality-of-life scores were signifi- Current Author Addresses: Drs. Pandya and Morrow: University of
Rochester Cancer Center, 601 Elmwood Avenue, Box 704, Roch-
cantly better in the oral clonidine group at 4 and 8 ester, NY 14642.
weeks, although the between-group difference was Dr. Raubertas: Department of Biostatistics, University of Roch-
small and the median did not differ. Figure 2 clearly ester, 601 Elmwood Avenue, Box 630, Rochester, NY 14642.
Dr. Flynn: 3800 Park Nicollet Boulevard, St. Louis Park, MN
shows that the response to both clonidine and pla- 55416-2699.
cebo varied considerably. Some women experienced Dr. Hynes: 929 North St. Frances, Wichita, KS 67201-1358.
no benefit from clonidine, and others reported sub- Dr. Rosenbluth: Northern New Jersey Cancer Center, 5 Summit
Avenue, Hackensack, NJ 07601-1992.
stantial improvement with placebo. Large placebo- Dr. Kirshner: 1000 East Genesee Street, Suite 400, Syracuse, NY
controlled studies are needed to ascertain the effec- 13210-1853.
tiveness of any putative agent in the control of hot Dr. Pierce: 1003 South Fifth Street, Second Floor, Tacoma, WA
98405-4210.
flashes, as evidenced by the placebo effect in this Dr. Dragalin: 1250 South Collegeville Road, Collegeville, PA
study. Baseline frequency of hot flashes did not 19426-0989.
influence response to clonidine treatment. Our Author Contributions: Conception and design: K.J. Pandya, P.J.
study was not designed for crossover treatment, and Flynn, R.F. Raubertas, J.J. Kirshner, G.R. Morrow.
therefore patient preference cannot be ascertained. Analysis and interpretation of the data: K.J. Pandya, R.F.
Raubertas, V. Dragalin, G.R. Morrow.
However, in our previous small pilot study, we ob- Drafting of the article: K.J. Pandya.
served a definite trend in favor of clonidine (9). Critical revision of the article for important intellectual con-
The toxicity of oral clonidine was low. Only 1 of tent: K.J. Pandya, R.F. Raubertas, J.J. Kirshner, G.R. Morrow.
Final approval of the article: K.J. Pandya, P.J. Flynn, J.J.
18 potential side effects differed significantly be- Kirshner, G.R. Morrow.
tween groups. In contrast, Goldberg and colleagues Provision of study materials or patients: K.J. Pandya, P.J.
(8) found statistically significant self-reported side Flynn, H.E. Hynes, R.J. Rosenbluth, J.J. Kirshner, H.I. Pierce.
Statistical expertise: R.F. Raubertas, V. Dragalin.
effects of dry mouth, constipation, and drowsiness Administrative, technical, or logistic support: G.R. Morrow.
with transdermal clonidine. Collection and assembly of data: G.R. Morrow.
Our study was designed for an intervention pe-
riod of 8 weeks, during which clonidine was found
to be effective. We believe that continued use of References
clonidine would provide clinical benefit, but we did 1. McKinlay SM, Jefferys M. The menopausal syndrome. Br J Prev Soc Med.
not examine its long-term effectiveness. Tamoxifen 1974;28:108-15.
2. Fisher B, Costantino J, Redmond C, Poisson R, Bowman D, Couture J,
and other selective estrogen receptor modulators et al. A randomized clinical trial evaluating tamoxifen in the treatment of
are now being tested in healthy postmenopausal patients with node-negative breast cancer who have estrogen-receptor-posi-
tive tumors. N Engl J Med. 1989;320:479-84.
women at high risk for breast cancer, and hot 3. Casper RF, Yen SS. Neuroendocrinology of menopausal flushes: an hypoth-
flashes are expected to be a side effect requiring esis of flush mechanism. Clin Endocrinol (Oxf). 1985;22:293-312.
4. Wren BG, Brown LB. A double-blind trial with clonidine and a placebo to
special attention. Oral clonidine may be helpful in treat hot flushes. Med J Aust. 1986;144:369-70.
5. Clayden JR, Bell JW, Pollard P. Menopausal flushing: double-blind trial of
that setting. We conclude that oral clonidine is use- a non-hormonal medication. Br Med J. 1974;1:409-12.
ful for controlling tamoxifen-induced hot flashes in 6. Edington RF, Chagnon JP, Steinberg WM. Clonidine (Dixarit) for meno-
pausal flushing. Can Med Assoc J. 1980;123:23-6.
postmenopausal women with breast cancer. 7. Laufer LR, Erlik Y, Meldrum DR, Judd HL. Effect of clonidine on hot
flashes in postmenopausal women. Obstet Gynecol. 1982;60:583-6.
8. Goldberg RM, Loprinzi CL, O’Fallon JR, Veeder MH, Miser AW, Mail-
From University of Rochester Cancer Center Community Clini- liard JA, et al. Transdermal clonidine for ameliorating tamoxifen-induced hot
cal Oncology Program Research Base, Rochester, New York; flashes. J Clin Oncol. 1994;12:155-589.
Metro-Minnesota Community Clinical Oncology Program, St. 9. Pandya KJ, Loughner J, Raubertas R, Bennett JM. A double blind placebo
Louis Park, Minnesota; Wichita Community Clinical Oncology controlled trial of clonidine for vasomotor symptoms in breast cancer patients
Program, Wichita, Kansas; Northern New Jersey Community on tamoxifen. Proceedings of the Annual Meeting of the American Society of
Clinical Oncology. 1990;9:340.
Clinical Oncology Program, Hackensack, New Jersey; Syracuse 10. Swartzman LC, Edelberg R, Kemmann E. The menopausal hot flush:
Hematology-Oncology Community Clinical Oncology Program, symptom reports and concomitant physiological changes. J Behav Med. 1990;
Syracuse, New York; and Northwest Community Clinical Oncol- 13:15-30.
ogy Program, Tacoma, Washington. 11. Morrow GR. Clinical trials in psychosocial medicine: methodological and
statistical considerations. Part III. Assessing measurement techniques in psy-
chosocial oncology. Cancer Treat Rep. 1980;64:451-6.
Acknowledgments: The authors thank the following additional 12. Erlik Y, Meldrum DR, Lagasse LD, Judd HL. Effect of megestrol acetate on
affiliates of the University of Rochester Cancer Center Commu- flushing and bone metabolism in post-menopausal women. Maturitas. 1981;
nity Clinical Oncology Program Research Base: Ernest Franklin, 3:167-72.
MD; David K. King, MD; Kevin P. Mulvey, MD; Brian Issell, 13. Bullock JL, Massey FM, Gambrell RD Jr. Use of medroxyprogesterone
acetate to prevent menopausal symptoms. Obstet Gynecol. 1975;46:165-8.
MD; Raymond S. Lord, MD; Richard H. Wheeler, MD; John 14. Albrecht BH, Schiff I, Tulchinsky D, Ryan KJ. Objective evidence that
Roberts, MD; and Paul Weiden, MD. placebo and oral medroxyprogesterone acetate therapy diminish menopausal
vasomotor flushes. Am J Obstet Gynecol. 1981;139:631-5.
Grant Support: By National Cancer Institute (CA37420). 15. Loprinzi CL, Michalak JC, Quella SK, O’Fallon JR, Hatfield AK, Neli-
mark RA, et al. Megestrol acetate for the prevention of hot flashes. N Engl
J Med. 1994;331:347-52.
Requests for Single Reprints: Kishan J. Pandya, MD, University of 16. Barton DL, Loprinzi CL, Quella SK, Sloan JA, Veeder MH, Egner JR, et
Rochester Cancer Center, 601 Elmwood Ave, Box 704, Roches- al. Prospective evaluation of vitamin E for hot flashes in breast cancer survi-
ter, NY 14642. vors. J Clin Oncol. 1998;16:495-500.

16 May 2000 • Annals of Internal Medicine • Volume 132 • Number 10 793

Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19956/ by a University of California San Diego User on 01/16/2017