OUP user menu

Gender differences in the long QT syndrome: effects of β-adrenoceptor blockade

Chantal E Conrath, Arthur A.M Wilde, Rosalie J.E Jongbloed, Mariëlle Alders, Irene M van Langen, J Peter van Tintelen, Pieter A Doevendans, Tobias Opthof
DOI: http://dx.doi.org/10.1016/S0008-6363(01)00477-1 770-776 First published online: 15 February 2002


Background: Gender differences have been reported in patients with the congenital long QT syndrome (LQTS). We analyzed whether electrocardiographic differences existed in females, males, girls and boys in response to β-adrenoceptor blockade. Methods: 12-lead ECGs before and during β-adrenoceptor blockade were collected in 87 genotyped LQTS patients (48 women, 14 men, 12 girls and 13 boys). Up to three QTc intervals were determined in each lead of the ECG. V4 was used for QT/QTc analysis. Difference between longest and shortest QT interval was taken as a measure for dispersion of QT intervals. Results: (1) Adult males had the greatest shortening of the QTc interval upon treatment with β-adrenoceptor blockade. During treatment, adult males with LQTS1 (mutation in the KCNQ1 gene, affecting IKs current) were found to have shorter QTc intervals than adult females; this difference did not exist in LQTS2 patients (mutation in the HERG gene, affecting IKr current). (2) Female LQTS2 patients had a 50% larger dispersion than female LQTS1 patients both before and during treatment. (3) Adult male LQTS1 patients constitute the only patient group with a marked decrease in QTc intervals and dispersion associated with a 100% efficacy of treatment in response to β-adrenoceptor blockade. Conclusions: These findings indicate that, in addition to underlying differences in repolarization between men and women, cardiac electrophysiological responses to β-adrenoceptor blockade can be modulated by gender-related factors.

  • Adrenergic (ant)agonists
  • ECG
  • Gender
  • Long QT syndrome
  • QT dispersion

Time for primary review 28 days.

1. Introduction

The QT interval is known to be influenced by gender [1–3]. Young boys and girls have similar QT interval durations. During puberty, the QT interval in boys shortens, leaving adult women with a longer QT interval than adult men [2]. In the congenital long QT syndrome (LQTS) adult women also have longer QT intervals than adult men [4,5]. Therefore they are more often clinically affected by this syndrome than men, in spite of the equal sex-distribution of the disease genotype. This is not restricted to electrocardiographic parameters. Female gender is an independent risk factor for cardiac events in patients with this syndrome [6]. Furthermore, there is a different time-dependent distribution of initial cardiac events [4,7]. In males, the probability of a first cardiac event by age 15 is higher than in females, and decreases after puberty, which it does not in females [7]. However, the first cardiac event is more often fatal in males than in females [7]. Thus, gender is a major determinant in the course and clinical manifestation of the long QT syndrome.

Women are more at risk than men of developing arrhythmias in response to QT prolonging drugs [8–10]. However, it is unknown whether there are differences in electrocardiographic response to β-adrenergic blocking agents between male and female carriers of long QT mutations. We have analyzed a genotyped LQTS population for sex linked differences in response to β-adrenoceptor blockade, with a distinction between adults and children. In addition, we analyzed whether these differences were specific for patients with mutations in the KCNQ1 gene (LQTS1) or the HERG gene (LQTS2), leading to changes in IKs and IKr current, respectively.

We found that during treatment with β-adrenoceptor blockade, men with LQTS1 have shorter QTc intervals than women; this was not seen in LQTS2 patients. Furthermore, female LQTS2 patients showed more dispersion than female LQTS1 patients. No difference in the genetic subgroups was found among adult males. Patients remaining symptomatic during treatment showed a positive relation between the amount of dispersion and the duration of the QTc interval, in line with basic concepts of arrhythmogenesis.

2. Methods

The study was performed according to a protocol approved by the local ethics committees in the academic hospitals of Utrecht, Amsterdam and Maastricht, and conforms with the principles outlined in the Declaration of Helsinki [11]. Written informed consent had been obtained from all patients. All genotyped LQTS1 and LQTS2 patients in whom no more than one mutation had been found and who were treated with β-adrenoceptor blockade were included in this study.

2.1. Genotype analysis

Patients were genotyped as previously described [12]. Briefly, genomic DNA was isolated from anticoagulated blood samples. Exons, encoding the complete sequence of the KCNQ1 and HERG (KCNH2) gene, were amplified by using PCR analysis (Perkin Elmer 9700 thermal cycler). Subsequently, amplicons were analyzed by SSCP analysis. Gels were run at 5 and 15°C, silver stained and air-dried. DNA fragments showing aberrant bands were purified (Qiagen PCR purification kit), sequenced by using the BigDye cycle sequencing kit (Applied Biosystems) and analyzed on an ABI-377 automatic sequencer (Applied Biosystems). In this study mutations were classified in both the gene involved (KCNQ1 or KCNH2) and the effect of the mutation, i.e. mutations leading to a stop-codon or a frameshift (truncating) and mutations leading to an amino acid substitution (missense). Three kinds of mutations were identified, i.e. LQTS1/missense (11 mutations), LQTS2/truncating (6 mutations) and LQTS2/missense (9 mutations).

2.2. Electrocardiography

In 12-lead ECGs before and during treatment up to three consecutive RR- and QT-intervals were measured in all leads. QTc intervals were calculated using Bazett's formula (QTc=QT/Embedded Image). All intervals were measured manually. The intersection of the line through steepest part of the downslope of the T-wave and the isoelectric line was defined the end of the T-wave. The end of the U wave was used in case fusion of T and U waves made distinction of the end of the T-wave impossible. In a minority of cases (mostly concerning the right precordial leads) the highest point of the second part of a biphasic T-wave was determined the end of the T-wave, based on T-wave morphology in subsequent leads. QT intervals were always determined similarly on the ECG before and the ECG during treatment in one patient. The difference between the longest and the shortest QT interval in any lead was used as a measure for dispersion.

2.3. Symptomatology

Symptomatology before and during treatment was determined. Patients who presented with cardiac arrest, registered torsades de pointes, a history of syncope, or, during treatment, presyncope that required alteration in their medication were defined as symptomatic. Follow-up during treatment was 5.5±5.7 years (average±S.D.). Three patient categories were defined based on clinical presentation: symptomatic before treatment and symptomatic during treatment (S–S), symptomatic before treatment and asymptomatic during treatment (S–A) and asymptomatic before and during treatment (A–A).

2.4. Statistical analysis

Differences between variances among the groups of boys, girls, males and females were tested by the F-test. In case of similarity of variances this was followed by one-way ANOVA before and during treatment. Changes in QT or QTc intervals following treatment were analyzed by the Student's t-test for paired observations to test differences from zero. Differences between changes in QT or QTc intervals between boys, girls, males and females were tested by one-way ANOVA. Values are given as mean±S.E.M. Levels of 0.05 or less were considered statistically significant.

3. Results

3.1. Change in QT and QTc intervals by β-adrenoceptor blockade in females, males, girls and boys

A total of 87 patients was included in this study: 48 women, 14 men, 12 girls and 13 boys (1–15 years). Fig. 1 shows that in boys (P<0.02) and females (P<0.001) QT intervals increased upon β-adrenoceptor blockade. Such an increase was not observed in males, whereas there was an insignificant increase in girls. Moreover, the difference between males and females was significant (ANOVA; P<0.05). Because RR intervals increased even more than QT intervals, QTc intervals decreased upon β-adrenoceptor blockade (Fig. 1). This decrease varied from 12±9.3 (NS) in girls to a largest value of 45±10.3 in males (P<0.001).

3.2. QTc intervals in men, women and children with LQTS1 and LQTS2

QTc intervals were similar in all groups (Fig. 2A). During treatment, adult males had significantly shorter QTc intervals than adult females in LQTS1, but not in LQTS2 patients (Fig. 2B; Table 1). Both males and females showed shortening of the QTc interval during treatment (compare Fig. 2A and B). No differences in response to β-adrenoceptor blockade existed between LQTS1 and LQTS2 in the four subgroups. In adult LQTS1 males shortening of the QTc interval upon treatment was 49±13.1 ms. In adult males with LQTS2 the QTc interval decreased by 42±15.9 ms during treatment. In both groups shortening was statistically significant (P<0.02 and P<0.05, respectively). Women with LQTS1 had a shortening of 37±16.7 ms and those with LQTS2 of 19±8.2 ms, which was in both groups significantly different from zero (P<0.05). In children the changes in QTc interval were less pronounced (compare Fig. 2A and B).

Fig. 1

Changes in QT and QTc intervals during β-adrenoceptor blockade are shown in boys, girls, males and females. #, Statistically significant change. In QT intervals this change was 20±7.2 (P<0.02), 18±10.7 (NS), −3±12.4 (NS), and 25±6.7 (P<0.001) in the four groups, respectively. The difference between males and females was statistically significant (ANOVA; P<0.05). Changes in QTc were −24±8.4 (P<0.02), −12±9.3 (NS), −45±10.3 (P<0.001) and −24±7.5 (P<0.01), respectively.

Fig. 2

QTc intervals in boys, girls, males and females with LQTS1 and LQTS2 are plotted before (A) and during (B) β-adrenoceptor blockade. See Table 1 for numerical data. Before treatment, QTc intervals were similar in all groups. During treatment, significantly shorter QTc intervals were found in male than in female LQTS1 patients, while no such difference was seen in LQTS2 patients. #, Statistical significance at P<0.01.

3.3. Gender differences in dispersion

In women dispersion differed significantly between LQTS1 and LQTS2 patients (Fig. 3). In men, however, the two genetic groups did not differ. Dispersion in females was 41±6.2 ms in the LQTS1 group and 62±3.7 ms in the LQTS2 group (P<0.01). The difference in dispersion between adult female LQTS1 and LQTS2 patients was also observed when girls were included in the female group. The larger dispersion in female LQTS2 patients proved to be independent of age (linear regression analysis; data not shown).

View this table:
Table 1

QTc intervals in females, males, girls, and boys with LQTS1 and LQTS2

nBefore treatmentDuring treatment
Fig. 3

Dispersion in QT intervals is shown in males and females before treatment. In women, LQTS2 patients have significantly more dispersion than LQTS1 patients, whereas in men no difference is found between LQTS1 and LQTS2 patients. See Table 1 for numerical data.

3.4. Symptomatology

Table 2 shows the efficacy of treatment in the LQTS1 and LQTS2 patient groups divided into boys, girls, males and females. Efficacy was defined as the percentage of S–A patients within the symptomatic (S–A+S–S) groups. Of course, the numbers are very small in most of the subgroups. In adults, the efficacy of treatment was 100% in male and 89% in female LQTS1 patients. In adult LQTS2 patients, efficacy of treatment was 67% in males and 70% in females.

View this table:
Table 2

Efficacy of treatment in females, males, girls, adn boys with LQTS1 and LQTS2


3.5. Symptomatology, QTc interval and dispersion

A positive relationship existed between the length of the QTc interval and the amount of dispersion in the adult S–S patients (seven women, one man) (Fig. 4). In addition, in this Fig. the type of LQTS is stated as well as the type of the mutation. Patients with a truncating LQTS2 mutation had the shortest QTc intervals and the least dispersion. Patients with missense mutations had larger dispersion and a longer QTc interval, and the only LQTS1 patient that remained symptomatic during treatment also had an extremely long QTc interval and a large dispersion (655 and 100 ms, respectively before treatment (see Fig. 4) and still 513 and 73 ms during treatment).

Fig. 4

Patients symptomatic during treatment show a positive relationship between the length of the QTc interval and amount of dispersion. 1, LQTS1; 2,LQTS2; Mis, missense mutation; trunc, truncating mutation. As indicated, LQTS2 patients with a truncating mutation have a relatively short QTc intervals and little dispersion. LQTS2 patients with missense mutations had a longer QTc interval and larger dispersion, as had the only (female) LQTS1 patient that remained symptomatic during treatment.

Fig. 5

Efficacy of treatment, (S–A/(S–A+(S–S))·100% is shown in the four groups subdivided into LQTS1 and LQTS2, in relation to the average QTc interval and the average dispersion for these patients in each group. B, boys; G, girls; M, males; F, females; 1 and 2, LQTS1 and LQTS2. The percentage represents the efficacy of treatment in each group (compare Table 2). A line separates the four groups with higher efficacy from the four groups with a lower efficacy. Shorter QTc intervals, less dispersion or a combination of both were related to a larger efficacy. In LQTS1 males, absence of symptoms during treatment was associated with a reduction in both dispersion and QTc duration. Efficacy of treatment tends to be larger in adult LQTS1 patients than in adult LQTS2 patients.

During treatment QTc intervals were shorter and dispersion was smaller in the four groups with the highest efficacy of treatment (89–100%) compared with the other four groups with lower efficacy (50–70%) (Fig. 5). In male LQTS1 patients both the reduction in QTc interval and the reduction in dispersion were the largest compared to all other patient groups.

4. Discussion

Patients with the congenital long QT syndrome demonstrate electrocardiographic differences during treatment with β-adrenoceptor blockade between adult males, females, and children (Figs. 1 and 2). Although no differences exist in QTc intervals before treatment among these groups, during treatment adult men in the LQTS1 group display the shortest QTc intervals (Fig. 2B). Women with LQTS2 have a significantly larger dispersion than women with LQTS1, both before (Fig. 3, Table 1) and during treatment (Table 1). In addition, in women that remained symptomatic during treatment, LQTS2 patients with a missense mutation had longer QTc intervals combined with a larger dispersion than those with a truncating mutation (Fig. 4). In patient groups with highest efficacy of treatment, QTc intervals were shorter and dispersion was smaller during β-adrenoceptor blockade (Fig. 5). This study is the first to show gender differences in electrocardiographic response to β-adrenergic blockade in LQTS patients.

4.1. QT(c) intervals in females, males, girls and boys

Differences in cardiac repolarization between men and women have been reported in healthy subjects and in LQTS patients [1–5,7]. Healthy women have longer QTc intervals and a higher heart rate. Stramba-Badiale et al. [3] showed QT differences to be more profound during low heart rates due to a steeper QT/RR relationship in women than in men. Also in the long QT syndrome women and children were reported to have longer QTc intervals than men [4,5], while female gender has been shown to be an independent risk factor for cardiac events [6].

To be able to understand differences in response to treatment in male and female subjects, it is necessary to understand the physiological background of gender differences in cardiac repolarization. Unfortunately, little is known about the influence of sex hormones on cardiac repolarization. In humans, androgens may shorten the QTc interval [2,4,5], whereas the effect of estrogens in humans is equivocal. Most studies addressing this issue have been performed in castrated rabbits treated with either estrogen or dihydrotestosterone (DHT). Pham et al. [13] showed that the duration of the first 30% of the action potential (APD30) was significantly shorter in castrated males than in castrated females, indicating a hormone-independent factor in differences in cardiac repolarization. An increase in QT interval was found in oophorectomized rabbits treated with estrogens compared to DHT, especially at long cycle lengths, due to a steeper QT/RR relationship [14]. In contrast, Drici et al. [15] found a lengthening of the QT interval in explanted hearts both from animals treated with male and from those treated with female sex hormones. In addition, they found a downregulation of the HK2 (human Kv1.5 [16]) and IsK (minK, a subunit to KCNQ1 that encodes the protein associated to IKs [17]) in both groups. The similar QT intervals in this study are not in conflict with other studies as QT intervals were measured at a short cycle length, while differences in QT interval are more pronounced at long cycle length. Besides a lower IK1, Liu et al. [18] identified a lower IKr current density in female than in male rabbits, which they held responsible for the steeper QT/RR relationship in female animals. These studies might form a rationale for the occurrence of longer QTc intervals in females. In our population, men and women had similar QTc intervals before treatment. This can be explained by the fact that we only included treated carriers. Upon treatment with β-adrenoceptor blockade, males had a greater shortening of the QTc interval than females. This was due to a shorter QTc during treatment in male than in female LQTS1 patients. This difference was not found in LQTS2 patients. Fig. 2A shows that before treatment women with LQTS1 had slightly larger QTc intervals than men. If findings in rabbits [18] hold true for the human species as well, this could be due to less IKr in the female human ventricle.

4.2. Gender differences in dispersion

We recently reported that LQTS2 patients in general display larger dispersion than LQTS1 patients [19]. However, no differences in dispersion were found between males with LQTS1 and LQTS2, whereas women with LQTS2 had significantly larger dispersion than those with LQTS1. A decrease of homogeneously distributed repolarizing IKr current may uncover a larger inhomogeneous contribution of IKs [20] to repolarization in LQTS2 patients, and thus lead to larger dispersion of action potential duration over the ventricular wall. Our data suggest that the inhomogeneous transmural distribution of IKs is larger in females than in males. This is supported by the fact that the QT/RR relationship is steeper in women than in men [5], which might be due to more IKs in the latter. To our knowledge there are no experimental data available to settle this issue. However, the fact that this difference between female LQTS1 and LQTS2 patients was similar in adult females and in girls and was independent of age excludes an important role for estrogens and suggests a hormone independent difference between the male and female heart.

In conclusion, gender differences exist in response to β-adrenoceptor blockade in patients with types 1 and 2 of the congenital long QT syndrome. During treatment males with LQTS1 have shorter QTc intervals than females and adult patients with LQTS2. Also treated females with LQTS1 exhibited less dispersion than those with LQTS2. Furthermore, an association was identified between efficacy of treatment and the relationship between QTc interval and dispersion. Studies to elucidate the underlying mechanisms of gender differences in response to β-adrenoceptor blockade will be needed to settle these issues.


We wish to thank G.C.M. Beaufort-Krol, M.Th.E. Bink-Boelkens and M.P. van den Berg, University Hospital Groningen, Groningen, L.J. Lubbers, Academic Medical Center, Amsterdam, R.F. Veldkamp, Medical Center Haaglanden, The Hague, F.A.L.E. Bracke, Catherina Hospital, Eindhoven, M. Witsenburg, Sophia Children Hospital, Rotterdam, and T.M. Hoorntje, R.N.W. Hauer and the late E.O. Robles de Medina, from the Department of Cardiology, University Medical Center, Utrecht for contributing patient data. This study was made possible by a grant of the Netherlands Heart Foundation (no. 95.014), and the Interuniversitary Cardiology Institute of the Netherlands (ICIN, project no. 27)


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
View Abstract