Discussion

Using this large, multicenter cohort with detailed manual review, we were able to robustly evaluate the effectiveness of the individual infection prevention strategies commonly used in clinical practice to reduce CIED infections. We found that many of these frequently applied interventions are ineffective and that the simplest strategies (those designed to limit bleeding risk and avoid implantations in patients with active infections) have the strongest potential to improve CIED infection outcomes. We also identified several strategies, such as prolonged use of antimicrobials after skin closure, that are both ineffective and increase harm, such as acute kidney injury and C. difficile infections.Reference Asundi, Stanislawski and Mehta¹⁹ Taken together and in combination with other high-quality studies,Reference Krahn, Longtin and Philippon⁹ these findings can be used to drive the development of evidence-based, cardiac-device–specific infection prevention protocols and guidelines.

Chlorhexidine skin cleaning is an intervention directly adopted from traditional surgical settings; it was also associated with reduced risk of postprocedural CIED infections. Chlorhexidine was likely effective in both settings because skin commensals introduced through incisions constitute the major microbiological culprits in both CIED procedures and conventional surgical settings.Reference Chen, Chen and Chen^20, Reference George, Leasure and Horstmanshof²¹ Similarly, the association we found between preprocedure MRSA status and CIED infection underscores the importance of skin commensals as drivers of procedure-related infections in this setting.

Preincisional antimicrobial prophylaxis is a mainstay of infection prevention in procedural settings. Although available evidence and surgical site infection prevention guidelines support only the use of a single preincisional dose,Reference de Oliveira, Martinelli and Nishioka^4, Reference Baddour, Epstein and Erickson^13, Reference Lee, Huang and Lin²² continuation of antibiotics for >1 day is common following electrophysiology procedures.Reference Mehrotra, Gupta and Strymish³ Prior studies demonstrate that prolonged post-procedural antibiotics following device placement are considered the standard of care by clinical electrophysiologists.Reference Basil, Lubitz and Noseworthy²³ In our cohort, there was no benefit to prolonging postprocedural antibiotics beyond a single preincisional dose; this finding is aligned with other studies in many surgical settings. Reducing postprocedural prescribing is an important stewardship effort with the potential to improve care by reducing preventable harms.Reference Bratzler^24, Reference Harbarth, Samore, Lichtenberg and Carmeli²⁵ Thus, implementing standardized protocols and guidelines to limit their use, or instituting a national quality metric, similar to the highly effective Surgical Care Improvement Project INF-3 measure, may be necessary to change this pervasive and potentially harmful practice.Reference Rosenberger, Politano and Sawyer^26, Reference Bratzler, Houck and Richards²⁷ Furthermore, identifying nonmodifiable risk factors that may trigger the application of ineffective interventions may be an area of focus for provider feedback. For example, revision procedures and generator changes are oft recognized as risk factors for infection. During unpublished qualitative interviews with electrophysiologists, we found that processes for revisions and generator replacements are more aggressive than prevention strategies applied to initial procedures. For example, during interviews, one provider reported, “Infections are a lot more common with generator changes… that’s why I extended the antibiotics for those patients, but I stop with the new devices.”

Overall, β-lactam antimicrobials appeared to be more effective for preventing CIED infections than vancomycin. Studies in other settings have generally favored β-lactam prophylaxisReference Bratzler, Dellinger and Olsen²⁸ as superior agents for Staphylococcus spp, the most common cause of CIED infections.Reference Bolon, Morlote and Weber²⁹ Another factor impacting vancomycin’s effectiveness as a prophylactic agent in this setting may be resource and environmental constraints that are present in the electrophysiology laboratory but not traditional operating rooms. Several providers reported during unpublished interviews that limited space and beds in the preprocedural area cause administration challenges with vancomycin, due to the need for prolonged infusion times with this antibiotic. These restraints prevent infusion of the antimicrobial prior to incision, thereby potentially limiting the effectiveness of the prophylaxis.

We also found that vancomycin prophylaxis also did not reduce CIED infection risk even after accounting for MRSA status; this finding is of particular importance given the strong association between preprocedure MRSA colonization status and postprocedural CIED infections. Our ability to explore the question of whether vancomycin was superior in the MRSA-colonized population was limited by the small number of patients tested for colonization at the time of their CIED procedure; however, infusion challenges described above would apply to all patients undergoing cardiac device procedures. In addition, supporting our findings is a recently published randomized controlled crossover trial measuring the impact of an antimicrobial prophylaxis bundle on CIED infection risk.Reference Krahn, Longtin and Philippon⁹ This randomized controlled trial found no additional benefit to adding vancomycin as part of a preincisional prophylaxis bundle.Reference Krahn, Longtin and Philippon⁹ Our work expands upon this study; we evaluated a broader variety of antibiotic regimens and also teased out the individual impact of different nonantibiotic bundle elements.Reference Krahn, Longtin and Philippon⁹ An additional factor that must be considered when evaluating the clinical utility of vancomycin as a prophylactic agent is its toxicity: prior work demonstrates a higher rate of acute kidney injury among patients who received vancomycin-based prophylaxis regimens compared to regimens without vancomcyin.Reference Asundi, Stanislawski and Mehta¹⁹ Taken together, these studies suggest that the most impactful way to minimize risk among MRSA-colonized patients might be an evidence-based skin antisepsis and decolonization bundle rather than a shift toward more widespread use of vancomycin as the primary agent for preincisional prophylaxis.

Some measures specifically targeted for CIED implantations (eg, the use of topical antibiotic washes for pocket irrigation and antibiotic-impregnated device envelopes) are widely used despite limited evidence supporting their effectiveness. Our analysis suggests that these topical administration strategies do not reduce CIED infections. In fact, we observed a trend toward an increased risk of infection among patients who received antimicrobial washes; this finding is congruent with prior studies evaluating this intervention that have demonstrated a lack of efficacy.Reference Krahn, Longtin and Philippon^9, Reference Kang, Liu and Liang^30, Reference Lakshmanadoss, Nuanez and Kutinsky³¹ Potential reasons why this strategy may not reduce infections include the risk of direct tissue toxicity from high doses of antimicrobials, increased risk of wound dehiscence, and the unpredictable pharmacodynamics of the diffusion of antimicrobials away from the device pocket.Reference Edin, Miclau, Lester, Lindsey and Dahners³² Another topical strategy, antimicrobial-impregnated TYRX envelopes, is specifically designed to reduce CIED infections and was approved for use in 2008. These absorbable and nonabsorbable mesh pockets release minocycline and rifampin in the generator pocket and have been associated with reduced infection risk in some observational studies.Reference Ahsan, Saberwal and Lambiase^7, Reference Henrikson, Sohail and Acosta^33, Reference Koerber, Turagam, Winterfield, Gautam and Gold³⁴ However, in our large, multicenter, national sample, the use of mesh pockets was not associated with reduced infection risk. This observation contrasts with a recently published randomized controlled trial supporting the efficacy of absorbable antibacterial envelopes.Reference Tarakji, Mittal and Kennergren³⁵ Reasons for this difference may include type of pocket evaluated; during the time frame of our study, both absorbable and nonabsorbable pockets may have been used and the newer pockets may have been more clinically effective than older designs. Another important difference is the 365-day infection outcome applied in the randomized controlled trial.

The high rate of wound complications preceding the development of CIED infections suggests that interventions that reduce the risk of wound complications (eg, those aimed at reducing bleeding) may lead to additional risk reduction. Consideration should be given to measures that reduce risk of hematoma, such as procedural checkpointsReference Strymish, Welch and Peralta⁵ to institute an INR cutoff of <1.5 with consideration to delay procedures if elevated, particularly in high-risk patients. The use of compression vests and compression dressings are other interventions that may be beneficial in patients who cannot safely stop their anticoagulants and/or antiplatelet agents.Reference Turagam, Nagarajan, Bartus, Makkar and Swarup^36, Reference Douketis³⁷ In addition, delaying procedures if a patient is febrile preprocedure may be a simple and effective risk reduction strategy.

This study has several limitations. The CART-EP application only captures a subset of voluntarily reporting VA electrophysiology procedures, and the sample may not be representative of all CIED procedures. In addition, the VA population may not be generalizable to non-VA populations, particularly to those with a higher proportion of female patients. Given the reliance on data available in VA records, adverse events occurring outside the VA system may not have been captured. However, prior studies have demonstrated that the majority of patients return to the closed VA healthcare system for subsequent care, somewhat mitigating this limitation.Reference Pindyck, Gupta and Strymish³⁸ In addition, we conducted a detailed manual review of scanned-in paper records including of outside facility records, so the vast majority of non-VA care was captured and reviewed. Although our cohort was strengthened by manual review of all patients, infection prevention efforts that were not documented would not have been measured; this would tend to bias our findings toward the null and was most likely to impact variables typically not captured in any medical record, such as type of skin antiseptic and receipt of shaving or clipping prior to the procedure. Furthermore, some variables may have only been intermittently documented (eg, chlorhexidine use), therefore underestimating the true scope of practice and the clinical effectiveness of this intervention.

Straightforward and easy-to-implement interventions should form the basis of standardized protocols and bundles (Fig. 3) to improve infection prevention in the electrophysiology laboratory. Although a variety of activities are commonly used in clinical practice, our findings highlight the principle that “less is more.”

Fig. 3.

Proposed Infection prevention bundle for CIED infections. *Nonmodifiable factors (eg, age, comorbidities, device type, and lead dislodgement) were omitted from this schematic because they are not often included in infection prevention initiatives. **These variables are potentially modifiable but may be difficult to target in the immediate preprocedural setting.