Early, Delayed and Late Cardiac Implantable Electronic Device Infections: Do the Timing of Onset and Pathogens Matter?

Infections involving cardiac implantable electronic devices (CIEDs) occur at different times after device-related procedures. The aim of this study was to investigate the timing of onset and factors influencing the occurrence of all types of CIED infections to identify the type of pathogen and to examine the long-term survival of patients with all types of CIED infections. We performed a post hoc analysis of the clinical data from 3344 patients who underwent transvenous lead extraction (TLE) at a single high-volume center between 2006 and 2020, including a group of 890 patients with CIED infections. The occurrence of pocket infection (PI), lead-related infective endocarditis (LRIE) and PI coexisting with LRIE (PI + LRIE) was assessed at the following time intervals: 0–12 months, 13–36 months and > 36 months since last CIED-related procedure. In the study group, there were 274 (30.79%) early infections, 266 (29.89%) delayed infections and 350 (39.32%) late infections. Pocket infection was the most common early complication (97; 39.43%), while LRIE was predominant over 36 months from the last CIED procedure (172; 54.09%). The most common early infections were PIs that were associated with the preceding CIED-related procedure. Late LRIE was most likely to occur in patients with intracardiac lead abrasion. The probability of early versus late LRIE was higher in patients with CoNS cultures. The timing of infection onset irrespective of its type does not affect long-term survival after transvenous lead extraction. The majority of infectious complications (69%) occur more than 12 months after the last CIED-related procedure. Early infections are probably associated with pocket contamination during CIED-related procedure, while delayed and late systemic infections are related to other lead-dependent factors (especially to intracardiac lead abrasion). Time to LRIE onset is associated with pathogen type. The timing of symptom onset does not affect long-term survival after TLE.


Introduction
Infectious complications related to cardiac implantable electronic devices (CIEDs) occur in 2.3-3.4% of CIED recipients [1][2][3]. Depending on the severity of the inflammatory process, there are three types of CIED infections: pocket infection (PI), lead-related infective endocarditis (LRIE) and PI coexisting with LRIE. Prior studies [4,5] have shown that the most common cause of LRIE is the spread of infection from the generator along the leads to 2 of 17 the endocardium. The analysis of the time relationship of infectious complications with the last CIED-related procedure may suggest the existence of a different mechanism involved in the pathogenesis of the different types of infections. Previous reports also suggested that early and late infections may be caused by different microorganisms, but research has been inconclusive. According to several studies, the most common pathogens causing both early and late infections are Staphylococcal species [6]. Another report found that there were more non-Staphylococcal infections in the late group compared to the early group [7], whereas, according to yet another study, Staphylococcus aureus was more likely to cause early than late infections [8]. In the present study, we assessed the occurrence of PI, PI + LRIE and LRIE at specified time intervals after the last revision procedure, and we also analyzed the types of pathogens and the long-term survival rate of patients with infectious complications.

Study Population
We performed a post hoc analysis of data from 3344 patients who underwent transvenous lead extraction (TLE) at a high-volume center (consisting of 3 branches in Lublin, Zamość, Radom in Poland) between 2006 and 2020. The clinical records from 890 (28.58%) patients with CIED infections were analyzed including 246 (27.64%) PI, 318 (35.73%) LRIE and 326 (36.63%) PI + LRIE cases. A total of 178 (24.78%) cases of possible endocarditis were included in the LRIE and PI + LRIE groups.

Analysis of Infectious Complications in Specific Time Intervals
Infectious complications were divided into: early-occurring within 0-12 months after preceding CIED-related procedure; delayed-appearing in the period from 13 months to 36 months; and late-found over 36 months after revision procedure. Factors potentially associated with the occurrence of PI, LRIE and PI + LRIE were analyzed at each time interval taking into account the clinical data (age, gender, concomitant diseases) and the device-related factors (type of system, number of leads, lead dwell time, lead-related abnormalities, procedures before TLE), as well as long-term survival after TLE (mean follow-up 1826 ± 1381 days). A microbiological analysis of cultures from device pockets, extracted leads and blood was also performed, assessing the type of pathogens at each time interval. Cultures from the removed leads were obtained only from patients without local pocket infection.

Definitions
A pocket infection was defined as the presence of local warmth, erythema, edema and pain or discharge from the device pocket or an erosion or impending erosion of the device according to the European Society of Cardiology (ESC) guidelines [9].
Lead-related infective endocarditis was diagnosed using modified Duke criteria according to the 2015 ESC guidelines for the management of infective endocarditis [9] and the European Heart Rhythm Association international consensus document [10]. The diagnosis of LRIE was definite in the presence of two major criteria or one major criterion and three minor criteria. Possible LRIE was diagnosed when the presence of one major criterion and one minor criterion or three minor criteria was confirmed.
Transvenous lead extraction procedure was defined according to the Heart Rhythm Society (HRS) and the European Heart Rhythm Association (EHRA) statements [11][12][13].
The term CIED-related procedure (revision procedure) was used to define the following types of procedures performed before infection: generator replacement, system upgrading or downgrading, placement of an additional lead and prior transvenous lead extraction.
Intracardiac lead abrasion was defined as outer insulation macroscopic damage, located in the intracardiac portion of the lead, usually in the first 15-20 cm from the tip, with visible discoloration, frequently with conductor externalization and often with purulent discharge [14].

Statistical Analysis
Distribution of all continuous variables was evaluated by the Shapiro-Wilk test. The distribution was non-parametric. More of them were expressed as means with standard deviation, wherein values of NYHA FC class and number of previous CIED-related procedures were expressed as median with IQR. All were compared with the ANOVA Kruskal-Wallis test. If there was a statistically significant difference in ANOVA Kruskal-Wallis test, the Mann-Whitney U test was used to compare all individual variables with each other. Categorical data are presented as absolute numbers and percentages and were compared using the Kruskal-Wallis ANOVA test. If there was a statistically significant difference in the Kruskal-Wallis ANOVA test, the χ 2 test with Yates correction was used to compare all individual variables with each other.
The patients were subdivided into groups based on temporal onset and type of infection related to last CIED procedure for comparison within the disease entity and between groups at a given time interval. Characteristics are marked with capital letters E (early), D (delayed) or L (late) in the case of statistically significant differences within the disease entity or Arabic numerals in the case of statistically significant differences between disease entities. Cox regression analysis was performed to identify the variables associated with infection. The survival time was defined as the period elapsing between the last CIED-related procedure (including initial CIED implantation) for noninfectious indications and TLE. The analysis was performed for each infectious complication (PI, PI + LRIE and LRIE) separately at each time interval (0-12. 13-36. >36 months): 0-12 monthsdata from all the 3114 patients were included; 13-36 months-patients who reached the study endpoints (infection) 12 months since last CIED-related procedure, but patients with noninfectious indications for TLE with follow-up <12 months were excluded from analysis; >36 months-patients who reached the study endpoints (infection) 36 months since last CIED-related procedure, but patients with noninfectious indications for TLE with follow-up <36 months were excluded from analysis.
To determine which factors have an impact on time to infection onset, univariable linear logistic regression was used to evaluate all the data from Tables 1 and 2. The variables having p values < 0.05 in univariable regression analysis were included in multivariable analysis. Logistic regression analysis included the data from patients with infection occurring 12 months and >36 months since last CIED-related procedure for noninfectious indications.
For each type of infectious complication, the Kaplan-Meyer curves were used to analyze the event-free survival between groups depending on temporal onset of infection, and the differences were tested for significance by the log-rank test.
A p value < 0.05 was considered statistically significant. Lack of statistical significance was described as NS (nonsignificant).

Results
In the population of 890 patients who underwent transvenous lead extraction due to CIED infection, there were 274 (30.79%) early infections (0-12 months since last CIEDrelated procedure), 266 (29.89%) delayed infections (13-36 months) and 350 (39.32%) late infections (>36 months). At 12 months after a CIED procedure, the distribution of the infectious complications was as follows: 97 (39.43%) PI, 117 (35.89%) PI + LRIE and 60 (18.87%) LRIE (p < 0.001 vs. PI and vs. PI + LRIE). More than 50% of PI, more than 60% of PI + LRIE and more than 80% of LRIE occurred at a later time after a CIED-related procedure (more than 12 months). Late isolated lead-related infective endocarditis was the dominant subgroup of all CIED infections (172 cases). The fewest LRIE cases were found in the early infection group, while the highest number was in the late infection group (54.09%). The opposite tendency was seen in patients with isolated pocket infection and with PI + LRIE; in these groups, device infection most commonly occurred 12 months since last CIED procedure (Table 1, Figure 1). CIED-cardiac implantable electronic device, LRIE-lead related infective endocarditis, PI-pocket infection.   Detailed analysis of clinical factors in individual types of infection in the studied time intervals showed that late LRIEs were more common in younger patients and in patients with heart failure with a higher NYHA class compared to late PI and late PI + LRIE (2 (IQR 1-3) vs. 2 (IQR 1-2) vs. 2 (IQR 1-2); p = 0.006).
There were significantly more women in the delayed LRIE group compared to the delayed PI and PI + LRIE groups (p = 0.044). Patients with LRIE more often had diabetes (27.36% vs. 22.92 vs. 22.70 in the PI and PI + LRIE groups; p = 0.013) and higher creatinine levels compared to the PI and PI + LRIE groups (1.59 ± 1.29 vs. 1.38 ± 1.09 vs. 1.29 ± 1.00; p < 0.001). These differences were significant in all analyzed time intervals.
Detailed analysis of the remaining clinical factors revealed non-specific differences between the different types of infection in the time periods studied.
With regard to device-related factors and type and timing of infection, early infections, compared to late ones, were more common in patients with high voltage (HV) lead-significant differences were found in both the PI and LRIE groups (32.99% vs. 17.72%, p = 0.009, and 41.67% vs. 26.74%, p = 0.026, respectively). Early, delayed and late LRIEs were more common in patients with the presence of lead in the coronary sinus.
Time-dependent analysis showed a greater number of leads in patients with early infections, especially in early PI and early PI + LRIE compared to late infections of this type (p= 0.026 and p < 0.001, respectively). Similarly, there was a tendency for early infection in patients with abandoned leads, especially in the early PI + LRIE group compared to delayed and late infections of this type (27.35% vs. 14.55% vs. 8.08%; p < 0.001).
The number of previous device-related procedures was lowest in patients with LRIE (1.79 ± 1.11 vs. 1.98 ± 1.13 in the PI + LRIE group vs. 2.09 ± 1.14 in the PI group; p = 0.004). Previous CIED-related procedures were on average significantly more common in the early PI rather than early LRIE group (2 (IQR 2-3) vs. 2 (IQR 1-2); p = 0.002). Moreover, in all types of infections, the number of revision procedures was lower in patients with late rather than early infections.
Dwell time of the oldest lead per patient during the last procedure before TLE was shorter in patients with late infections compared to early infections of all types. Late PI and late LRIE more often occurred in patients with longer lead dwell time during TLE compared to early infections of this type (113.7 ± 73.77 vs. 89.59 ± 58.71 months; p < 0.001 and 104.3 ± 61.07 vs. 67.33 ± 64.63 months; p < 0.001).
Patients with late LRIE were more likely to have excessively elongated loops of the leads (9.30% vs. 5.06% in patients with late PI vs. 3.03% in patients with late PI + LRIE; p = 0.006). Similarly, the phenomenon of abrasion of leads was more often found in delayed and late LRIE compared to late PI and late PI + LRIE (33.72% vs. 27.14% vs. 12.73%; p= 0.002).
Late infections of all types, especially late LRIE, were less likely to be associated with the preceding revision procedure (65.12% late LRIE occurred after primary implantation of CIED, without any preceding procedure), while the most early infections, especially early PIs (59.79%) were found after unit replacement ( Table 2).
Analysis of cultures from the pocket, leads or blood showed that coagulase-negative Staphylococci (CoNS) were more often found in early vs. delayed and late LRIE (6.67 vs. 5.81% vs. 1.16%; p = 0.05). Staphylococcus epidermidis was most common in patients with early PI + LRIE compared to early PI (28.21% vs. 12.37%; p= 0.018) and less common in delayed LRIE compared to delayed PI + LRIE (16.28% vs. 31.82%; p= 0.034). This pathogen was also less frequent in patients with late PI + LRIE compared to early and delayed infections of this type (14.14% vs. 28.21% vs. 31.82; p = 0.009).
Staphylococcus aureus was most often found in patients with PI + LRIE, especially in the group of late PI + LRIE compared to late isolated LRIE (18.18% vs. 7.56%; p = 0.012).
Negative cultures were observed more often in patients with early PI compared to LRIE + PI (35.77% vs. 20.51%; p = 0.014) and in the group of delayed LRIE compared to delayed LRIE + PI (32.56% vs. 16.36%; p = 0.01) ( Table 3).      Multivariate Cox regression analysis showed that the main risk factor of isolated pocket infections in all analyzed time intervals was any CIED-related procedure before TLE, and the probability of PI occurrence was highest in the early period after the revision procedure (HR = 7.350; 5.507-10.865 95%CI; p < 0.001). The probability of PI in all analyzed time intervals also increased with the age of patients during the last procedure before TLE. The risk of late PI was lower in women (HR = 0.468; 0.284-0.769 95%CI; p = 0.003).
The probability of developing PI + LRIE increased with any CIED-related procedure, especially in the early period since the revision procedure (HR = 8.388; 6.070-11.591 95%CI; p < 0.001). The risk of PI + LRIE in all analyzed time intervals also increased with the patient's age during the preceding procedure and was higher in patients with abandoned leads. The probability of early and delayed PI + LRIE increased in patients with a higher creatinine level. The risk of early, delayed and late PI+ LRIE was lower in patients with a shorter dwell time of the oldest lead. The probability of the delayed and late PI + LRIE was also lower in females and in patients that used long-term anticoagulation therapy.
The probability of occurrence of the isolated LRIE in all time intervals increased with any CIED-related procedure and was highest in the early period (HR-5.733. 95%CI: 3.662-8.975; p < 0.001). The risk of early, delayed and late LRIE was also higher in patients with abrasion of the leads. The probability of late LRIE increased in patients with a higher NYHA class and higher creatinine level (HR = 1.490. 95%CI: 1.123-1.978; p = 0.006 and HR = 1.270. 95%CI 1.137-1.418; p < 0.001). The risk of isolated LRIE in the late period after last procedure was also higher in patients with lead in the coronary sinus and with abandoned lead (HR = 2.322. 95%CI 1.471-3.664; p < 0.001 and HR = 2.210. 95%CI 1.213-4.029; p = 0.010). The probability of early LRIE was higher in patients with the leads on both sides of the chest (HR = 3.752; 95%CI 1.036-13.588; p = 0.044).
The probability of LRIE in all analyzed time intervals decreased in patients with a shorter dwell time of the oldest lead (Table 4).  The comparative analysis of the risk of early versus late infections showed that the probability of late infections increased in patients with a higher NYHA class (OR = 1.024. 95% CI 1.008-1.038; p = 0.002). Risk of early infectious complications increased with the age of patients and with the body mass index (BMI) (OR = 1.654. 95% CI 1.213-1.256, p < 0.001, and OR = 1.053. 95% CI 1.004-1.103, p = 0.032, respectively).
Patients with more leads were more likely to develop late PI + LRIE, but the tendency was only seen in the univariate analysis, whereas an increased risk of early versus late PI + LRIE was observed after any CIED-related procedure and in patients with the presence of Staphylococcus epidermidis in cultures (OR = 2.326. 95% CI 0.993-5.435; p = 0.051).
The probability of early LRIE was higher in patients with HV leads (OR = 3.367. 95% CI 1.401-8.130; p = 0.006), with older leads during last CIED-related procedure (OR = 2.326 95% CI 0.993-5.435; p = 0.051) and with the presence of CoNS in cultures (OR = 10.00 95% CI 1.429-71.43; p = 0.020) ( Table 5).  At long-term follow-up after TLE (1826± 1381 days), the probability of event-free survival at all time intervals was highest in patients with PI (about 45% after 10 years) and lowest in those with LRIE (about 35% after 10 years) (Figure 2A). Detailed analysis of each group: isolated PI ( Figure 2B), PI + LRIE ( Figure 2C) and isolated LRIE ( Figure 2D) showed no significant differences in survival rates between successive time intervals. At long-term follow-up after TLE (1826± 1381 days), the probability of event-free survival at all time intervals was highest in patients with PI (about 45% after 10 years) and lowest in those with LRIE (about 35% after 10 years) (Figure 2A). Detailed analysis of each group: isolated PI ( Figure 2B), PI + LRIE ( Figure 2C) and isolated LRIE ( Figure 2D) showed no significant differences in survival rates between successive time intervals.

Discussion
Infectious complications in patients with cardiac implantable electronic devices occur at various times after initial placement and appear to be frequently associated with previous device procedures. Prior studies have shown that rates of infection were two to five times higher after revision procedures than after initial implantation (initial 0.5-0.8%, revision 1-4% at 1 to 3 years of follow-up) [15][16][17]. In the present study, the rate of cardiac device infection after a CIED-related procedure was only slightly higher than that observed after initial implantation (54.61% vs. 45.39%), and the most common infectious

Discussion
Infectious complications in patients with cardiac implantable electronic devices occur at various times after initial placement and appear to be frequently associated with previous device procedures. Prior studies have shown that rates of infection were two to five times higher after revision procedures than after initial implantation (initial 0.5-0.8%, revision 1-4% at 1 to 3 years of follow-up) [15][16][17]. In the present study, the rate of cardiac device infection after a CIED-related procedure was only slightly higher than that observed after initial implantation (54.61% vs. 45.39%), and the most common infectious complications were found up to 12 months after generator replacement procedures. Multi-year analysis of 62 infection cases documented similar results: 33 (53.2%) infections occurred after initial implantation, 18 (29.0%) after the second procedure and 11 (17.7%) after the third and more procedures [18]. Previous reports comparing early and late CIED infections showed that the symptoms of pocket infection (erythema, pain, swelling, warmth and pus or drainage from the pocket) were likely to occur 6 to 12 months since last device procedure, whereas the signs of systemic symptoms predominated in late infections [7,[19][20][21]. The current study confirmed a relatively high incidence of PI in the early period after the preceding procedure, while LRIE was more frequent in more than 36 months after the last CIEDrelated procedure. However, the present study demonstrated that a high rate of PI also occurred at >12 months after any revision procedure (over 50% of infectious complications, including 23.26% of PI occurring at >60 months after last CIED procedure). The same tendency was already observed in one previous report [7]; however, the factors that may favor the development of late pocket infections without a direct relationship with previous CIED procedure have not been identified so far. It was postulated that the risk of late PI was higher in patients with a too-shallow generator location and, consequently, progressive erosion of the pocket, as well as with the identification of less virulent pathogens, mainly coagulase-negative Staphylococci [7].
The spread of an infection from the device pocket along the leads to the endocardium is, according to most investigators, the most likely cause of LRIE (systemic infection) [7,[19][20][21][22][23][24][25][26][27]. In the present study, PI + LRIE occurred at a similar rate at all time intervals, whereas isolated LRIE was more evident over an extended period of time (>50% of LRIE cases occurred at >5 years after last procedure). Moreover, the present study revealed that the probability of early versus late LRIE increased with the presence of CoNS, which supports the concept of bacteria spreading from the infected pocket along the leads to the endocardium. However, such a coincidence was confirmed only for early LRIE, especially coexisting with PI, whereas late systemic infections, most common, were associated with the process of intracardiac lead abrasion. The influence of lead abrasion on the development of lead-related infective endocarditis was described in our previous reports [14,[28][29][30], whereas the current study confirmed that the pathogenesis of late isolated LRIE was different and probably not directly related to bacterial contamination of the pocket. This explanation may be supported by the higher incidence of Staphylococcus aureus in patients with delayed systemic infections in this study. Moreover, we demonstrated that the development of late LRIE was influenced by clinical factors such as heart failure and renal failure. Probably, reduced immunity increases the risk of infection, and externalized conductors due to abrasion-related breach of the insulation facilitate the migration of pathogens along the lead and formation of biofilm.
Long-term follow-up of patients undergoing TLE for infectious indications showed high mortality in patients with systemic infections. The relationship between time to infection onset and long-term survival has not been established. Due to the high mortality rate of patients with CIED-related infections, it is important to minimize infectious complications. In order to reduce the risk of infection, antibacterial envelopes, leadless pacemakers and subcutaneous ICD systems are increasingly used. Current research confirms the good effects of new technologies. The use of leadless pacemakers allows for the complete elimination of the risk of pocket infection. Recent studies confirm that by improving the leadless pacemaker implantation technique, electrical stability is ensured, which reduces the need for reintervention [31]. In patients requiring ICD implantation, the use of subcutaneous s-ICD eliminates the occurrence of LRIE, and according to recent studies, the risk of devicerelated complications related to s-ICD during the 2-year follow-up period is relatively rare [32].

Limitations
This is a post hoc analysis, and therefore no randomization was performed. Our findings are limited by the ability to analyze only the patient population undergoing TLE.

Conclusions
Most cardiac device infections occur more than 12 months after the preceding CIEDrelated procedure. The occurrence of various types of infections at different time intervals after last revision procedure is associated with specific risk factors. Early PI is more often associated with last CIED procedure, whereas the risk of LRIE, especially late after last revision procedure, is higher in patients with intracardiac lead abrasion. The pathogenesis of PI + LRIE is probably related to the contamination of device pocket with low-virulent pathogens (most often CoNS) during initial implantation or during previous CIED procedure. The timing of infection onset irrespective of its type does not affect long-term survival after transvenous lead extraction. The persistently high mortality of patients with CIED-related infections requires the search for methods to minimize the risk of infectious complications.