Customize your JAMA Network experience by selecting one or more topics from the list below.
Drici M, Knollmann BC, Wang W, Woosley RL. Cardiac Actions of Erythromycin: Influence of Female Sex. JAMA. 1998;280(20):1774–1776. doi:10.1001/jama.280.20.1774
From Georgetown University Medical Center, Washington, DC. Dr Drici is now with the University of Nice-Sophia Antipolis, Nice, France.
Context.— Erythromycin is a widely used antibiotic that infrequently causes QT-prolongation
and torsades de pointes cardiac arrhythmias. For antiarrhythmic drugs, women
are at a higher risk for these cardiac arrhythmias, but few other classes
of drugs have been studied.
Objectives.— To determine whether female sex is a risk factor for cardiac arrhythmias
associated with erythromycin, and if this can be correlated with in vitro
measurements of the QT-response to erythromycin in male and female rabbit
Design.— Food and Drug Administration (FDA) MEDWATCH database analysis and in
Main Outcome Measures.— Cardiac arrhythmia reports associated with erythromycin from 1970 until
1996 classified by patient sex and age, and effect of female sex on erythromycin-induced
QT-prolongation in isolated perfused rabbit hearts.
Results.— We observed a sex difference in cardiac arrhythmias associated with
administration of erythromycin. A total of 346 cases were found in the FDA
database: 201 females (58%), 110 males (32%), and 35 unspecified (10%). Forty-nine
were life-threatening ventricular arrhythmias and deaths directly related
to intravenous erythromycin lactobionate: 33 women (67%) and 16 men (33%)
(P = .03). During the same period, no sex imbalance
was present in the prescription pattern for intravenous erythromycin lacobionate
(men 47%, women 49%, unspecified 4%). Perfusion with erythromycin caused significantly
greater QT-prolongation in female rabbit hearts (mean [SD], 11.8% [2.3%])
than in male hearts (6.9% [2.1%]; P = .03).
Conclusions.— As has been shown in reports of antiarrhythmic drugs, we found a female
predominance in the FDA reports of erythromycin-associated cardiac arrhythmias.
Based on in vitro experiments, a sex difference in cardiac repolarization
response to erythromycin is a potential contributing factor.
ERYTHROMYCIN is a widely used macrolide antibiotic for which lengthening
of the rate-corrected QT interval (QTc) on the electrocardiogram (ECG) has
been reported.1,2 A prolonged
QTc interval on the ECG is a feature associated with drug-induced torsades
de pointes (TdP), an arrhythmia reported in patients treated with erythromycin.3-6
Men between the ages of 15 and 60 years have a shorter QTc interval
than women of the same age.7,8
Our previously reported experiments in ovariectomized rabbits have suggested
that an action of sex hormones on myocardial tissue is responsible for this
difference.9 Likewise, this is supported by
the strong sex difference in the clinical occurrence of TdP with antiarrhythmic
drugs. A preponderance of the cases (70%) occur in women compared with the
predicted incidence of 40%.10,11
To evaluate the potential influence of sex and age on the cardiac adverse
effects of erythromycin, we searched the Food and Drug Administration's (FDA's)
MEDWATCH Spontaneous Reporting System (SRS) for such events. The overall pattern
of erythromycin usage was estimated from the National Disease and Therapeutic
Index (NDTI, IMS Health Inc, Plymouth Meeting, Pa). We then used an in vitro
model to examine a potential sex difference in the effect of erythromycin
on the QTc interval of isolated perfused rabbit hearts.
The FDA SRS database consists of voluntary reports of adverse events
associated with drugs marketed in the United States that are submitted primarily
by health care professionals. A search of the FDA SRS database was conducted
to identify adverse drug reaction reports associated with erythromycin products
between 1970 and 1996 under the Freedom of Information Act. The search was
limited to cardiovascular arrhythmias, such as ventricular fibrillation, TdP,
heart arrest, tachycardia, bradycardia, atrioventricular blocks, extrasystoles,
QT interval prolongation, ECG abnormalities, and combinations of these terms.
The reports were classified according to the sex and age of the patient.
To estimate the sex and age distribution of erythromycin use in the
United States, drug prescription audits included in the NDTI were analyzed
for the years 1991 to 1996. The NDTI is a survey designed to provide statistical
information about the patterns and treatment of disease encountered in office-based
practice in the United States. Prescriptions for erythromycin lactobionate
filled at Georgetown University Medical Center, Washington, DC, were analyzed
Experiments were conducted in accordance with the guidelines of the
Georgetown University Animal Care and Use Committee. Five male and 5 female
albino New Zealand rabbits were studied. The Langendorff method for isolated
heart perfusion and measurement of QT intervals was used as described previously.9 Briefly, the hearts were excised, mounted in a modified
Langendorff perfusion apparatus, paced at a constant cycle length of 400 milliseconds,
and perfused with modified Tyrode solution at 37°C. Increasing concentrations
of erythromycin lactobionate (Abbott Laboratories Inc, Abbott Park, Ill) at
1, 10, and 100 µmol/L were added for successive periods of 40 minutes
to study the QTc response at concentrations similar to the range reached in
plasma during clinical therapy (10-50 µmol/L).1,12
Three standard ECG leads were recorded and the QTc interval measured by 2
blinded investigators (M.D.D. and W.-X.W.), as previously described.9 After equilibration of the preparation for baseline
measurements, erythromycin was added to the perfusion for 40 minutes. At baseline
and again at the end of the erythromycin perfusion of 100 µmol/L, the
pacing cycle length was switched abruptly to 1000 milliseconds for 5 minutes,
then back to 400 milliseconds, to determine the magnitude of the effect of
erythromycin at different pacing rates (rate dependence).
Sex differences of adverse drug reaction reports were tested by the χ2 statistic ([Observed Values−Expected Values]2/Expected
Values), with P<.05 considered significant. Expected
values and odds ratios were calculated on the basis of observed adverse events
in each sex and the sex distribution from the NDTI database. Confidence intervals
for the odds ratios were based on central limit theorem estimates. The effect
of erythromycin on QT intervals was tested by analysis of variance, with Bonferroni-Dunn
From 1970 through 1996, 346 incidents of cardiac arrhythmias involving
erythromycin products were reported via the FDA SRS. Sex was not reported
in 35 cases. Overall, significantly more cases of arrhythmias occurred in
women compared with men (201 vs 110, P=2.5×10−7). This was especially apparent between the ages of 10 and 59
years (Figure 1, A).
A potential cause of the higher incidence of arrhythmias in women could
have been a greater number of prescriptions of erythromycin for women. We
therefore analyzed the age and sex distribution in the NDTI. From 1991 to
1996, a total of 78 million prescriptions for erythromycin products
were written: 56% for women, 41% for men, and 3% for patients whose sex was
not specified. We then used the age-specific sex difference in erythromycin
prescriptions to estimate any underlying sex difference in erythromycin exposure. Figure 1, B shows the odds ratios for female
sex as a risk factor for erythromycin-related arrhythmias, calculated for
each age decade. The overall increased risk for women remained statistically
significant, with female sex a risk factor, especially during the reproductive
Because of the nature of the FDA SRS database of adverse drug events,
duplicate entries can occur, and often the causal relationship between drug
exposure and adverse reaction is poorly defined. To validate the results of
our initial search strategy, we reviewed all reports related to the intravenous
administration of erythromycin and selected only life-threatening arrhythmias
(ventricular tachycardia, TdP, ventricular fibrillation, and cardiac arrest)
directly associated with the infusion of erythromycin lactobionate. Forty-nine
cases were documented. Again, predominantly female patients were affected:
33 female cases (67%) and 16 male cases (33%) were described (P=.03). Torsades de pointes or ventricular tachycardia in the presence
of a long QTc interval was specifically reported in 36 cases (24 females and
12 males, P=.05). The sex difference was most apparent
between the ages of 15 and 50 years (Figure
1, C). The prescription pattern for the intravenous formulation
erythromycin lactobionate, which accounted for one third of all reported arrhythmias,
was more balanced between male and female patients than oral erythromycin.
One million erythromycin lactobionate prescriptions were reported, 47% to
male patients and 49% to female patients. Sex was unspecified in 4% of the
cases. Similarly, at Georgetown University Medical Center during the same
period erythromycin lactobionate was given to 846 patients, 49% male and 51%
We then studied the effect of erythromycin in isolated hearts from female
and male rabbits. At baseline, females had a QT interval (SD) of 240 (2) milliseconds
vs 232 (4) milliseconds in male rabbit hearts (P=.16).
Perfusion of 1 and 10 µmol/L of erythromycin lactobionate increased
the QT interval to 244 (3) milliseconds and 250 (3) milliseconds in females,
and did not change the QT interval in males (228  milliseconds and 229
 milliseconds; P = .01 between sexes). The perfusion
with 100 µmol/L of erythromycin lactobionate increased the QT interval
in both sexes with a resulting QT duration at steady state of 266 (6) milliseconds
in females vs 243 (4) milliseconds in males (P =
.01). This represented a prolongation of 11.8% (2.3%) in female hearts compared
with 6.9% (2.1%) in male hearts (Figure 2). The overall drug-effect relationship was significantly different
in males and females (P = .03), with males requiring
more than 10 times the concentration of erythromycin as females to produce
the same QT interval prolongation.
The sex difference was even more prominent at the slower pacing rate
studied. Under baseline conditions, a switch to a slower pacing rate prolonged
the QT interval similarly in both sexes: 18.8% (1.8%) in females vs 19.3%
(1.2%) in males (P=.82). With 100 µmol/L of
erythromycin lactobionate, the same change from rapid to slow pacing produced
more pronounced QT interval prolongation in females (26.1% [2.3%]) than in
males (22.8% [1.6%]; P = .003). Perfusion with erythromycin
caused significantly greater QT-prolongation in female rabbit hearts than
in male rabbit hearts (P = .03).
This study demonstrates a sex difference in the reported adverse cardiac
effects associated with erythromycin, with a preponderance in women (about
two thirds), even when corrected for differences in the number of prescriptions
written for each sex. This was particularly true for arrhythmias associated
with the intravenous formulation of erythromycin, erythromycin lactobionate.
Our findings also demonstrate a sex difference in erythromycin's ability to
induce QT interval prolongation in isolated perfused rabbit hearts. We showed
that erythromycin displays a sex-related reverse rate dependence, ie, a greater
QT interval lengthening at slow heart rate (60/min) in females. Those 2 results
together with our prior demonstration of a male sex hormone–induced
blunting of cardiac response to quinidine's effect on repolarization9 might contribute to the strong sex difference that
we observed in the retrospective reports of cardiovascular adverse effects
induced by erythromycin.
There were considerably more cardiac arrhythmias reported in women between
the ages of 15 and 50 years than in men. This is the same age when the sex
difference in QTc interval is greatest between men and women.8
Similar observations have been made in patients with familial long QT syndrome.
Men exhibit shorter QTc intervals than both women and children,13
and after puberty women are at a higher risk for arrhythmic events than men.14
The greater risk of women to develop TdP during therapy with potassium
channel blockers has been recognized with antiarrhythmic drugs for several
years.10,15 Because of the differences
in body size and the fact that men and women are often given the same dose,
higher serum concentrations could be responsible for the increased risk of
TdP in women. However, this is not necessarily the case for erythromycin,
because the rate of erythromycin metabolism is higher in women.12
Our current results in isolated male and female rabbit hearts indicate that
the 2 sexes have a quantitative difference in response to the same drug concentration.
The data from the FDA SRS database must be interpreted with caution
because of the inherent limitations of a spontaneous voluntary reporting system.
Possibly only as few as 1 in 10 or 1 in 100 serious adverse events are actually
reported to the FDA. However, we are not aware of any factors that would introduce
a sex bias in voluntary reporting. Based on our prescription survey data,
we conclude that at least for intravenous erythromycin the prescription pattern
had no sex bias.
In summary, we have found that the clinical case reports of cardiac
arrhythmias associated with erythromycin display a significant sex difference
with a predominance in women. Experimentally, normal male and female rabbit
hearts have a different profile of QT interval lengthening during perfusion
with erythromycin lactobionate, with a blunted response to erythromycin in
male hearts. Therefore, we hypothesize that an increased sensitivity to erythromycin
in women could facilitate, especially at slow heart rates, the induction of
TdP ventricular arrhythmias.
Create a personal account or sign in to: