Sex differences in selection of pacemakers: retrospective observational study

Abstract

Objective: To evaluate the effect of patients’ sex on selection of pacemakers.

Design: Retrospective univariate and multivariate analysis of a large database.

Setting: German central pacemaker register.

Subjects: Records collected at the register for 1992 and 1993 (n=31 913), covering 64% of all implantations in Germany.

Main outcome measure: Probability of receiving a single chamber, dual chamber, or rate responsive pacemaker in relation to sex.

Results: Univariate analysis showed that women were more likely to receive single chamber pacemakers and less likely to receive dual chamber or rate responsive systems than men. After demographic and clinical variables were controlled for, women were still more likely to receive a single chamber system (atrial pacing: odds ratio 0.89, 95% confidence interval 0.74 to 1.07; ventricular pacing: 0.85, 0.80 to 0.92) and less likely to receive a dual chamber (1.20, 1.12 to 1.30) or a rate responsive system (1.26, 1.17 to 1.37) than men.

Conclusions: The data suggest sex differences in the selection of a pacemaker system which cannot be explained by the underlying cardiac disorder. Further research is needed to evaluate why guidelines for implanting pacemakers are not better adhered to.

Key messages

• Use of pacemakers varies despite guidelines, and the reasons for this are unclear

• In this study women were more likely to receive single chamber pacemakers and less likely to receive dual chamber and rate responsive pacemakers than men

• Demographic and clinical variables cannot fully explain these differences

• Prospective studies are needed to evaluate the effect of sex and other non-medical variables on the selection of pacemakers

Introduction

Pacemakers are the standard treatment for symptomatic bradyarrhythmia. Professional societies have issued guidelines for the implantation of cardiac pacemakers.^1–3 Although these guidelines have all been similar, considerable differences have been reported in the frequency of implantation of pacemakers and in the system selected.⁴ Relatively little is known about the reasons for these differences.⁵ There is some evidence that patients’ sex might play a part in clinical decision making.^6–8 We used a large database to analyse whether sex can explain differences in the selection of pacemaker systems.

Subjects and methods

Doctors who perform implants transfer information available from the European pacemaker patient identification card to the German central pacemaker register on a voluntary basis. Customs’ statistics and manufacturers’ sales figures suggest that the register comprises 64% of all implantations in Germany. Of the 880 German hospitals in which implants are done (number estimated by a survey), 634 (72%) sent their reports to the register.⁹ Our analysis is based on complete data for the years 1992 and 1993. We excluded cases that had classification errors (for example, non-existent categories) from the total of 15 914 patients in 1992 and 15 999 patients in 1993. Table 1 gives the characteristics of the patients.

Table 1.

Characteristics of patients receiving pacemaker by sex^*

	Men				Women
	No (%)	Mean age	Median age		No (%)	Mean age	Median age	P value†
Age (years)	n=15 648				n=15 110
Mean (SD)		73.8 (12.1)			77.6 (10.9)			<0.001
Median (interquartile range)			73.6 (65.8-80.8)				77.4 (70.4-82.4)
Type of pacemaker	n=15 884				n=15 389
Single chamber atrial	429 (2.7)	67.5	68.3		492 (3.2)	70.9	72.8	<0.005
Single chamber ventricular	10 499 (66.1)	73.2	76.4		11 065 (71.9)	75.8	78.7	<0.001
Dual chamber	4892 (30.8)	67.7	70.1		3770 (24.5)	69.8	72.4	<0.001
Rate responsive‡	4273 (26.9)	68.1	70.5		3186 (20.7)	70.0	72.8	<0.001
Symptoms	n=12 005				n=11859
Syncope	4154 (34.6)	71.8	74.5		4554 (38.4)	75.4	78.3	<0.001
Dizziness	3229 (26.9)	72.1	73.7		3368 (28.4)	74.3	77.0	<0.01
Bradycardia	2389 (19.9)	71.8	73.3		2063 (17.4)	73.9	77.3	<0.001
Heart failure	708 (5.9)	73.1	76.2		629 (5.3)	75.8	78.7	<0.05
Electrocardiographic result	n=11 874				n=11 788
Atrioventricular block	4346 (36.6)	71.5	74.0		3902 (33.1)	74.1	78.0	<0.001
Sick sinus syndrome	3930 (33.1)	72.0	73.7		4644 (39.4)	74.8	77.2	<0.001
Atrial fibrillation/flutter	2209 (18.6)	73.4	75.8		2181 (18.5)	76.1	78.7	0.814
Aetiology	n=1 397				n=11 263
Ischaemic	3932 (34.5)	73.7	75.5		3627 (32.2)	76.1	78.5	<0.001
Fibrosis	878 (7.7)	71.7	73.8		969 (8.6)	74.9	77.5	<0.01
Carotid sinus syndrome	524 (4.6)	71.0	72.4		338 (3.0)	75.9	78.3	<0.001

^*Categories with small numbers and cases labelled unknown are not included in the table.  

^†Age analysed by t test, all other variables by χ² test.  

^‡Percentage of the above pacemaker systems with rate responsive function. 

Statistical analysis

We examined the pooled data for 1992 and 1993 according to patient characteristics (age, sex, clinical symptoms, underlying disease, electrocardiographic findings), type of pacemaker (single chamber, dual chamber), and rate responsive systems. The χ² test was used to analyse discrete variables and Student’s t test used for continuous variables.

For logistic regression we defined the type of pacemaker as the dependent variable and age, sex, clinical symptoms, underlying disease, and electrocardiographic findings as independent variables. Explanatory variables for the selection of a specific pacemaker were first determined in a univariate logistic regression. We then used a multivariate model adjusting for age, clinical symptoms, underlying disease, and electrocardiographic findings to determine the effect of sex on choice of pacemaker. Cases with more than one missing value were excluded from univariate regression, and only complete datasets were used in the multivariate regression. All data were processed with sas 6.08 software. Α P value <0.05 was considered significant. All tests were two tailed.

Results

Table 1 shows that single chamber systems were implanted in a higher proportion of women (atrial 492 (3.2%), ventricular 11 065 (71.9%)) than men (atrial 429 (2.7%), ventricular 10 499 (66.1%)). In contrast, men received more dual chamber (4892 (30.8%)) and rate responsive systems (4273 (26.9%)) than women (3770 (24.5%), 3186 (20.7%)). Sex differences were also found for various clinical variables.

In the univariate analysis, where 21 858 patients could be included (68.5% of the sample), age, sex, several clinical symptoms, underlying disease, and electrocardiographic findings were significantly associated with the pacemaker system selected for implantation (table 2). After the effects of age, clinical symptoms, underlying disease, and electrocardiographic findings had been adjusted for in a multivariate analysis (n=16 289 (51.0% of the data)), women were still more likely to receive a ventricular single chamber pacemaker and less likely to receive a dual chamber or rate responsive system than men (table 3).

Table 2.

Univariate analysis of variables determining selection of pacemaker in 21 858 patients^*

	Type of pacemaker (odds ratio (95% CI))
	Single chamber atrial	Single chamber ventricular	Dual chamber	Rate responsive
Age	0.84 (0.80 to 0.88)	1.35 (1.33 to 1.39)	0.75 (0.70 to 0.80)	0.80 (0.78 to 0.82)
Sex	0.84 (0.71 to 0.98)	0.76 (0.72 to 0.81)	1.38 (1.30 to 1.47)	1.38 (1.29 to 1.47)
Symptoms:
Syncope	1.04 (0.87 to 1.21)	1.15 (1.08 to 1.23)	0.85 (0.79 to 0.91)	0.65 (0.60 to 0.69)
Dizziness	1.64 (1.39 to 1.90)	0.78 (0.73 to 0.83)	1.20 (1.13 to 1.29)	1.19 (1.10 to 1.27)
Bradycardia	0.74 (0.58 to 0.91)	1.11 (1.03 to 1.20)	0.93 (0.85 to 1.01)	1.09 (1.00 to 1.18)
Heart failure	0.15 (0.07 to 0.25)	1.09 (0.95 to 1.23)	1.03 (0.90 to 1.18)	1.54 (1.35 to 1.75)
Electrocardiographic results:
Atrioventricular block	0.06 (0.04 to 0.09)†	0.35 (0.32 to 0.38)	3.85 (3.57 to 4.16)	0.93 (0.87 to 1.01)
Sick sinus syndrome	13.29 (11.11 to 15.42)	0.89 (0.84 to 0.94)	0.80 (0.75 to 0.85)	1.19 (1.12 to 1.27)
Atrial fibrillation/flutter	0.11 (0.07 to 0.17)†	14.28 (12.48 to 16.08)	0.06 (0.05 to 0.07)‡	1.01 (0.92 to 1.08)
Aetiology:
Ischaemic	0.71 (0.59 to 0.83)	1.43 (1.33 to 1.52)	0.71 (0.66 to 0.77)	0.68 (0.63 to 0.74)
Fibrosis	0.03 (0.02 to 0.04)	0.81 (0.73 to 0.90)	1.28 (1.15 to 1.43)	1.04 (0.93 to 1.16)
Carotid sinus syndrome	0.59 (0.35 to 0.88)	2.63 (2.13 to 3.17)	0.37 (0.30 to 0.45)	0.34 (0.27 to 0.43)

^*Categories used in logistic regression were age: older patients/younger patients (10 year intervals); sex: men/women; and symptoms, electrocardiography, and aetiology: yes/no.  

^†n<25.  

^‡n=119—cases with rare indications, low degree of disorder (atrioventricular block), or intermittent disorder. 

Table 3.

Variables determining selection of pacemaker in multivariate logistic regression on 16 289 patients^*

Type of pacemaker	No (%) receiving treatment		Multivariate adjusted odds ratio* (95% CI)
	Men	Women
Single chamber atrial	2226 (2.8)	260 (3.2)	0.89 (0.74 to 1.07)
Single chamber ventricular	5680 (69.6)	6069 (74.6)	0.85 (0.80 to 0.92)
Dual chamber	2252 (27.6)	1802 (22.1)	1.20 (1.12 to 1.30)
Rate responsive†	1852 (22.7)	1482 (18.2)	1.26 (1.17 to 1.37)

^*Adjusted for age, electrocardiographic results, symptoms, and aetiology; the ratios are men/women.  

^† Number of above pacemaker systems with rate responsive function. 

Discussion

Analysis of data from over 15 000 patients suggests a sex bias in choice of a pacemaker system. Women were more likely to receive single chamber systems and less likely to receive dual chamber or rate responsive systems than men. Can these findings be explained by differences in the underlying cardiac disorders or demographic data? In our cohort a higher proportion of men presented with an atrioventricular block than women. It is generally accepted that dual chamber pacemakers achieve better haemodynamic results than single chamber systems in atrioventricular block.²,¹⁰ Women, on the other hand, had a higher frequency of sinus node dysfunction. In this disorder a single chamber system often seems to be sufficient, although several authors have found that dual chamber pacemakers produce better outcomes in terms of haemodynamics, subjective symptoms, the development of atrial fibrillation, and prognosis.¹¹ Finally, cardiovascular diseases occur at a later age in women than in men.¹² In our cohort women were on average 3.8 years older than men. Doctors generally implant single chamber pacemakers in elderly patients rather than dual chamber systems.¹³

Several studies of factors influencing cardiovascular interventions showed that sex was no longer a determinant once demographic and clinical variables had been adjusted for.¹⁴,¹⁵ In our study, however, even after we controlled for demographic and various clinical variables sex remained independently associated with the selection of a pacemaker system. Our results agree with two retrospective studies in the United States in which women were found to receive a dual chamber system less frequently than men.¹⁶,¹⁷ The clinical importance of the suggested undertreatment of women with dual chamber and rate responsive pacemakers is not easy to evaluate. In addition to the advantages of dual chamber pacemakers in patients with atrioventricular block and sinus node dysfunction mentioned above, rate responsive systems have been shown to offer haemodynamic advantages over fixed rate systems in patients with chronotropic incompetence receiving ventricular single chamber pacing or dual chamber pacing.¹⁸ Although there is evidence that patients treated by advanced pacing have a better quality of life,¹⁹ it is not known whether this improvement is equal in men and women.

Is there sex discrimination?

What other reasons could there be for doctors deciding in favour of a single chamber pacemaker in women? Firstly, there are some “soft” indications for implanting pacemakers (class II indications in the American College of Cardiology/American Heart Association guidelines² and corresponding recommendations in many countries, including Germany) Furthermore, guidelines are not always unanimously adhered to in clinical practice.⁵,²⁰ Doctors are known to behave differently towards men and women as far as both diagnostic and therapeutic strategies are concerned.²¹ Doctors seeing women with “soft” indications may tend to implant single chamber pacemakers whereas they choose dual chamber for men. Some of the “hard” indications may also be being neglected in women. Women often present their symptoms differently from men.²² They are more likely to receive the same treatment as men if they present their symptoms as men do.²³ Finally, we found some published evidence that women sometimes reject sophisticated care in favour of more simple treatments. They may therefore choose not to have dual chamber systems.²⁴

The database we used represents two thirds of the implantations performed in Germany. Since the percentage of reporting hospitals slightly exceeds the percentage of reported implantations, hospitals with a lower frequency of implantations may be overrepresented. However, for this majority of hospitals (72%) sex differences were present in selection of pacemakers. Sex was an independent determinant of choice of pacemaker, with women receiving roughly 20% fewer rate responsive and dual chamber systems then men. Missing data meant that we could include only 16 289 of the 31 913 records in the multivariate logistic regression analysis, and we could not control our data for variables such as left ventricular function, intermittent dysrhythmia, or multimorbidity that might have contributed to the differences found. In addition, our results do not necessarily apply to other countries. However, a similar sex bias has been shown to be likely in the United States at least.¹⁶,¹⁷ Study limitations and the retrospective design of our analysis do not allow a definite explanation for the sex bias. Prospective studies that include clinical endpoints such as survival or quality of life are needed to investigate this difference in more detail and reveal its potential implications.