Real-world evidence on the association between cardiac implantable electronic device infection and all-cause mortality

Abstract

Aims

An infection following cardiac implantable electronic device (CIED) procedure is a serious complication, but its association with all-cause mortality is inconsistent across observational studies. To quantify the association between CIED infection and all-cause mortality in a large, contemporary cohort from New South Wales, Australia.

Methods and results

This retrospective cohort study used linked hospital and mortality data and included all patients aged >18 years who underwent a CIED procedure between July 2017 and September 2022. Cardiac implantable electronic device infection was defined by the presence of relevant diagnosis codes. Cox regression to estimate adjusted hazard ratios (aHRs) with 95% confidence intervals (CIs) for the association of CIED infection with mortality, at 1-year, and at the end of follow-up, with CIED infection included as a time-dependent variable, and other potential risk factors for mortality included as fixed covariates. We followed 37,750 patients with CIED procedures {36% female, mean age [standard deviation (SD)] 75.8 [12.7] years}, and 487 (1.3%) CIED infections were identified. We observed 5771 (15.3%) deaths during an average follow-up of 25.2 (SD 16.8) months. Compared with no infection group, patients with CIED infection had a higher Kaplan-Meier mortality rate (19.4 vs. 6.8%) and adjusted hazard of mortality (aHR 2.73, 95% CI 2.10−3.54) at 12 months post-procedure. These differences were attenuated but still remained significant at the end of follow-up (aHR 1.83, 95% CI 1.52−2.19).

Conclusion

In a complete, state–wide cohort of CIED patients, infection was associated with higher risks of both short–term and long–term mortality.

Graphical Abstract

Graphical Abstract.

Background

The global proliferation of cardiac implantable electronic devices (CIEDs) is accompanied by a corresponding rise in the incidence of device-related infections.¹ According to a recent comprehensive review, all-cause mortality following CIED infection could reach up to 35% over up to 5.5 years of long-term follow-up.² The prospective WRAP-IT trial (Worldwide Randomized Antibiotic Envelope Infection Prevention Trial) showed a more than three-fold increase in all-cause mortality at 12-month follow-up among patients with CIED infection compared with those without infection.³ A population-wide claims–based study of United States Medicare beneficiaries indicated a significantly increased mortality risk post-CIED infection, which lasted for up to 3 years after pacemaker infections and 2 years for single- and dual-chamber implantable cardioverter defibrillator (ICD) infections.⁴ Another observational study focusing on patients whose device was explanted (and thereby excluding any patients who died prior to explantation) has reported no heightened mortality risk with CIED infection.⁵ Additionally, a population-wide study from Denmark with an average follow-up of 5.2 years found no increased risk of mortality among the infected group as compared to their uninfected counterparts but had no information on the proportion of device explantation.⁶ Given these findings, our study seeks to quantify the association between CIED infection and all-cause mortality in a broad, contemporary cohort from New South Wales (NSW), Australia’s largest and most populous state.

In this retrospective cohort study, we utilized person-level data linked from hospital records [the NSW Admitted Patient Data Collection (APDC)] and mortality registries (NSW Registry of Births, Deaths, and Marriages). The study received ethical approval from the NSW Population and Health Services Research (HREC/18/CIPHS/56). We included patients aged over 18 years who underwent a CIED procedure during the period from July 2017 to September 2022 for public hospitals, and until June 2021 for private hospitals. These patients were identified using relevant Australian Classification of Health Interventions (ACHI)⁷ codes in primary and secondary procedure fields. We classified CIED device types into four categories: permanent pacemakers (PPM), ICD, cardiac resynchronization therapy with pacemaker (CRT-P), and cardiac resynchronization therapy with defibrillator (CRT-D). Furthermore, CIED procedures were categorized as de novo, revision, or replacement (including upgrades). Cardiac implantable electronic device infection was defined by the presence of relevant diagnosis code [T82.71: infection and inflammatory reaction due to electronic cardiac device, including infected pacemaker or defibrillator (pocket) or infection due to lead or electrodes] in any hospital record following a CIED procedure.

Patients were followed from the date of their CIED procedure until either their death or the end of the follow-up period, whichever occurred first. The cumulative mortality rate among CIED infection and no infection groups was presented using Kaplan–Meier plots. We used Cox regression to estimate adjusted hazard ratios (aHR) with 95% confidence intervals (CIs) for the association of CIED infection with mortality, at 1-year and at the end of follow-up, with CIED infection included as a time-dependent variable and other potential risk factors for mortality included as fixed covariates.

Of 37 750 patients who underwent CIED procedures {36% female, mean age [standard deviation (SD)] 75.8 [12.7] years}, 487 (1.3%) CIED infections were identified. Nearly one-third (32.7%) of the identified infections occurred within the first week of a CIED procedure. Half (50.1%) of patients with CIED infection had their device removed, and 80.3% of these removals occurred within 7 days of the patient’s first hospital admission with a diagnosis of infection. Table 1 shows the characteristics of patients undergoing a CIED procedure according to whether they had a CIED infection or not. Compared to those who did not have CIED infection, those with CIED infection were more likely to be younger (mean age 72.1 years vs. 75.9 years), have devices with CRT capabilities (15.6% vs. 8.1%), replacement or revision procedures (38.2% vs. 24.9%), and medical comorbidities, and have the CIED procedure at a public hospital (62.0% vs. 57.0%), noting that the prevalence of diabetes, congestive heart failure, COPD, chronic kidney disease, and obesity was higher in public hospital patients. They also had more previous CIED procedures (40.0% vs. 20.0%) and previous CIED infections (13.1% vs. 0.6%) (Table 1).

Table 1.

Selected characteristics of patients with and without infection following cardiac implantable electronic device procedure

Factors	Number (%)		P-value
	No infection (n = 37 263)	Infection (n = 487)
Mean age in years (SD)	75.9 (12.7)	72.1 (15.4)	<0.001
Sex			0.054
Male	23 673 (63.5)	330 (67.8)
Female	13 590 (36.5)	157 (32.2)
Device type			<0.001
PPM	27 574 (74.0)	286 (58.7)
ICD	6699 (18.0)	125 (25.7)
CRT-P	1587 (4.3)	30 (6.2)
CRT-D	1403 (3.8)	46 (9.4)
Procedure type			<0.001
De-novo	27 995 (75.1)	301 (61.8)
Revision	563 (1.5)	16 (3.3)
Replacementc	8705 (23.4)	170 (34.9)
Previous CIED procedurea			<0.001
No	29 809 (80.0)	292 (60.0)
Yes	7454 (20.0)	195 (40.0)
Previous CIED infectiona			<0.001
No	37 058 (99.4)	423 (86.9)
Yes	205 (0.6)	64 (13.1)
Emergency admission			0.885
No	25 212 (67.7)	331 (68.0)
Yes	12 051 (32.3)	156 (32.0)
Private insurance			0.006
No	16 970 (45.5)	252 (51.7)
Yes	20 293 (54.5)	235 (48.3)
Medical comorbiditiesb
Diabetes	11 417 (30.6)	170 (34.9)	0.042
Hypertension	29 225 (78.4)	389 (79.9)	0.440
Coronary artery disease	20 203 (54.2)	307 (63.0)	<0.001
Congestive heart failure	14 741 (39.6)	285 (58.5)	<0.001
Atrial fibrillation	14 819 (39.8)	234 (48.0)	<0.001
Peripheral vascular disease	9523 (25.6)	190 (39.0)	<0.001
Stroke	3565 (9.6)	71 (14.6)	<0.001
COPD	6093 (16.4)	109 (22.4)	<0.001
Sleep apnoea	1434 (3.8)	32 (6.6)	0.002
Chronic kidney disease	8962 (24.1)	182 (37.4)	<0.001
Obesity	6777 (18.2)	123 (25.3)	<0.001
Cardiomyopathy	3878 (10.4)	97 (19.9)	<0.001
Concomitant cardiac surgery			0.557
No	36 062 (96.8)	469 (96.3)
Yes	1201 (3.2)	18 (3.7)
Had CABG surgery in previous 12 months			0.214
No	36 868 (98.9)	479 (98.4)
Yes	395 (1.1)	8 (1.6)
Had valve surgery in previous 12 months			0.683
No	36 804 (98.8)	480 (98.6)
Yes	459 (1.2)	7 (1.4)
Hospital type			0.027
Public	21 245 (57.0)	302 (62.0)
Private	16 018 (43.0)	185 (38.0)

CABG, coronary artery bypass grafting; CIED, cardiac implantable electronic device; COPD, chronic obstructive pulmonary disease; CRT-D, cardiac resynchronization therapy defibrillator; CRT-P, cardiac resynchronization therapy pacemaker; HR, hazard ratio; ICD, implantable cardioverter defibrillator; PPM, pacemaker.

P-values are from chi-square test for categorical variables and two-sample t-test for continuous variables.

^aCalculated by looking into hospitalizations since 1 July 2001.

^bCalculated by looking into hospitalizations in the preceding 12 months.

^cReplacement includes upgrade.

A total of 5771 (15.3%) patients died during an average follow-up of 25.2 (SD 16.8) months. Compared to CIED patients without infection, patients with CIED infection had a higher Kaplan–Meier mortality rate (19.4% vs. 6.8%) and adjusted hazard of mortality (aHR 2.73, 95% CI, 2.10–3.54) at 12 months post-procedure (Figure 1). This increased hazard was attenuated but still remained significant at the end of follow-up (aHR 1.83, 95% CI 1.52–2.19).

Figure 1.

Kaplan–Meier all-cause mortality curves. Hazard ratios (HR) are estimated from Cox proportional hazard regression model with adjustment for age, sex, private insurance, CIED device type, CIED procedure type, previous CIED procedure, previous CIED infection, concomitant procedure, previous CABG or valve surgery, hospital funding type, and comorbidities including diabetes, hypertension, coronary artery disease, congestive heart failure, atrial fibrillation, peripheral vascular disease, stroke, COPD, sleep apnoea, chronic kidney disease, obesity, and cardiomyopathy.

In our comprehensive study involving NSW residents undergoing CIED procedures, we observed a strong association between CIED infection and elevated 12-month and longer term mortality, even after adjusting for patient and hospital characteristics. Our 12-month Kaplan–Meier mortality rates were slightly higher than, but comparable to, those reported by the WRAP-IT trial (16% and 5%, for patients with and without infection, respectively).³ Our risk-adjusted HRs were lower than, but within the confidence bounds of, those reported by WRAP-IT (HR 3.41 95% CI 1.81–6.41 at 12 months and HR 2.30 95% CI 1.29–4.07 over follow-up of up to 36 months).³ Importantly, our findings come from a whole-of-population cohort undergoing a full spectrum of CIED procedures, with around three-quarters of these being de novo PPM insertion. In contrast, the WRAP-IT trial included only high-risk patients undergoing CIED generator replacement or upgrade, CIED pocket or lead revision, or an initial CRT-D procedure.³ The pronounced difference in mortality during the early follow-up months between groups with and without CIED infection may be attributable to (i) complications linked to CIED infection such as sepsis, infective endocarditis, and embolism¹ and (ii) the interruption of CIED therapy, i.e. a period when the patient was without previously indicated device therapy.³ Once patients overcome the initial event of CIED infection and its subsequent complications, other causes of death likely start to exert a more profound influence over time. With longer-term follow-up, these non–device-related causes of death may attenuate the association but do not negate the initial significant impact of CIED infection on mortality. One limitation of our study is the use of a single diagnostic code to identify CIED infections which could encompass everything from a localized infection to bacteraemia and septic shock. This was evident in our results where despite device removal being a class I recommendation following a CIED infection,¹ only 50.1% of such devices were removed. This lower removal rate suggests that many CIED infections detected using the diagnostic code might have been localized, leading to more conservative management.

Our findings provide important real-world evidence to inform decisions about the potential benefits of interventions to reduce the risk of CIED infection, such as the use of absorbable antibiotic-eluting envelopes which has been found to be associated with a reduction in CIED infection in recent studies.^8–10

Funding

This study was funded by grants from the National Health and Medical Research Foundation of Australia (NHMRC) grant (#APP11447430) and Medtronic Australasia Pty Ltd.

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request.