Incidence of carditis and predictors of pacemaker implantation in patients hospitalized with Lyme disease

Background Lyme carditis, defined as direct infection of cardiac tissue by Borrelia bacteria, affects up to 10% of patients with Lyme disease. The most frequently reported clinical manifestation of Lyme carditis is cardiac conduction system disease. The goal of this study was to identify the incidence and predictors of permanent pacemaker implantation in patients hospitalized with Lyme disease. Methods A retrospective cohort analysis of the Nationwide Inpatient sample was performed to identify patients hospitalized with Lyme disease in the US between 2003 and 2014. Patients with Lyme carditis were defined as those hospitalized with Lyme disease who also had cardiac conduction disease, acute myocarditis, or acute pericarditis. Patients who already had pacemaker implants at the time of hospitalization (N = 310) were excluded from the Lyme carditis subgroup. The primary study outcome was permanent pacemaker implantation. Secondary outcomes included temporary cardiac pacing, permanent pacemaker implant, and in-hospital mortality. Results Of the 96,140 patients hospitalized with Lyme disease during the study period, 10,465 (11%) presented with Lyme carditis. Cardiac conduction system disease was present in 9,729 (93%) of patients with Lyme carditis. Permanent pacemaker implantation was performed in 1,033 patients (1% of all Lyme hospitalizations and 11% of patients with Lyme carditis-associated conduction system disease). Predictors of permanent pacemaker implantation included older age (OR: 1.06 per 1 year; 95% CI:1.05–1.07; P<0.001), complete heart block (OR: 21.5; 95% CI: 12.9–35.7; P<0.001), and sinoatrial node dysfunction (OR: 16.8; 95% CI: 8.7–32.6; P<0.001). In-hospital mortality rate was higher in patients with Lyme carditis (1.5%) than in patients without Lyme carditis (0.5%). Conclusions Approximately 11% of patients hospitalized with Lyme disease present with carditis, primarily in the form of cardiac conduction system disease. In this 12-year study, 1% of all hospitalized patients and 11% of those with Lyme-associated cardiac conduction system disease underwent permanent pacemaker implantation.

Introduction Lyme disease is the most common tick-borne disease in the United States, Canada, and Europe [1,2]. Previously published studies report that up to 10% of patients hospitalized with Lyme disease exhibit clinical manifestations of the infection of cardiac tissue by Borrelia bacteria [3][4][5][6][7][8][9]. These clinical manifestations are commonly referred to in aggregate as Lyme carditis and are identified during the acute disseminated phase of the infection in most patients. The most frequently reported manifestation of Lyme carditis is cardiac conduction system disease (CCD). Less frequently reported cardiac manifestations of Lyme disease include acute myopericarditis and dilated cardiomyopathy [1,2,5,6,10,11].
Atrioventricular (AV) block in the context of Lyme carditis can progress rapidly to complete heart block (CHB). Therefore, patients with Lyme disease who present with evidence of CCD are typically hospitalized and monitored continuously with cardiac telemetry. Although Lyme carditis-related CCD usually resolves with antibiotic therapy, there can be a lag between the initiation of antibiotics and recovery of cardiac conduction. Consequently, some patients require temporary cardiac pacing (TCP) support during the recovery of the cardiac conduction system [3,12,13]. Permanent pacemaker (PPM) implantation is reserved for symptomatic, advanced CCD that persists despite antibiotic treatment [3,10].
The objective of this study was to investigate the impact of Lyme carditis on the outcomes of patients hospitalized with Lyme disease in the US. Multiple outcomes were studied including mortality, heart failure, and PPM implantation.

Query of the National Inpatient Sample database
The data utilized in this study were obtained from the National Inpatient Sample (NIS), the largest collection of all-payer data on inpatient hospitalizations in the United States. The NIS database includes patient-and hospital-level data, including: patient demographics, primary and secondary diagnoses, interventions and procedures, and medical comorbidities (consistent with definitions set forth by the Agency for Healthcare Research and Quality (AHRQ)), and length of stay [14]. For this study, data were obtained from the NIS for the years 2003-2014. Because all NIS datasets are de-identified, this study was exempt from institutional review by the Human Research Committee.
Each unique data entry in the NIS database represents a single hospitalization with a principal discharge diagnosis, secondary diagnosis entries (maximum 29), and procedure entries (maximum 15) [14]. Patients diagnosed with Lyme disease were identified using the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) code 088.81.

Study design
In this study, a patient was defined as having Lyme carditis if they were admitted to an inpatient unit with a diagnosis of Lyme disease during the study period and were also diagnosed with CCD, acute myocarditis, or acute pericarditis. Demographic data were collected for all patients included in the study. Comorbid conditions were identified using measures provided by the AHRQ. Demographic and comorbidity data were used to calculate the Elixhauser comorbidity index for each patient included in the study [15,16]. Patients who underwent TCP or PPM implantation were identified using the appropriate procedure codes. Among patients with CCD, those who had undergone pacemaker implantation prior to the index hospitalization (N = 310) were excluded from this analysis.

Statistical analysis
Trend weight files were applied in all analyses to account for sampling methods used in the construction of the NIS dataset [17]. All analyses of trends were conducted with the nonparametric test for trends by Cuzick [18].
Baseline characteristics of patients hospitalized with Lyme disease during the study period are displayed in Table 1. Multivariable logistic regression analysis was utilized to identify predictors of patient outcomes, including Lyme carditis, TCP, PPM implantation, and in-hospital mortality. For all analyses, survey estimation was utilized to account for the complex survey design of the NIS database. In all comparisons, a P value of <0.05 was considered significant.
All analyses were completed using Stata/SE software version 14.1 (StataCorp LP, College Station, Texas).

Baseline characteristics
A total of 96,140 patients were hospitalized with Lyme disease during the study period (Fig 1). The baseline characteristics of patients hospitalized with Lyme disease in the United States between 2003 and 2014 are summarized in Table 1. There was a near-equal distribution of females and males (51% and 49%, respectively) in the study population. The median (IQR) age was 51 (34-66) years. Most patients in this study were white (88%).
The most prevalent comorbidity among patients hospitalized with Lyme disease was hypertension (32%). Other commonly reported comorbidities included chronic lung disease (12%) and diabetes mellitus (11%). A significant increase in the prevalence of all comorbidities except valvular disease was observed between 2003 and 2014 (Table 1). This increase was also evident in the Elixhauser score, as the number of patients with an Elixhauser score greater than 2 increased from 22% in 2003 to 41% in 2014 (P trend = 0.002).

Temporal and regional trends in Lyme disease-related hospitalizations
There was a significant increase in the number of patients hospitalized with Lyme disease during the study period, with a near-doubling from 4,801 in 2003 to 9,375 patients in 2014 (P trend = 0.003). A comparable difference was observed among patients with Lyme carditis ( Table 2). A majority of the hospitalizations for Lyme disease occurred in the Northeast region of the United States (61%). The South, Midwest, and West accounted for 22%, 13%, and 5% of hospitalizations, respectively.

Lyme carditis in patients hospitalized with Lyme disease
Patients with Lyme carditis constituted 11% (N = 10,465) of all patients hospitalized with Lyme disease during the study period (Fig 1). New CCD was the most common manifestation of Lyme carditis and was observed in 10% of patients hospitalized with Lyme disease during the study period (93% of patients with Lyme carditis). Acute myocarditis and acute pericarditis were rare and comprised 2.1% and 0.3% of all patients hospitalized Lyme disease, respectively.

Clinical manifestations of Lyme carditis in hospitalized patients
Among the 10,465 patients hospitalized with Lyme disease who also had evidence of Lyme carditis, 9,729 (93%) were diagnosed with CCD. Acute myocarditis was also reported in 2,002

Pacemaker implantation in patients with Lyme carditis
Outcomes of patients hospitalized with Lyme disease who also had evidence of Lyme carditis are summarized in Table 5. Of the 9,729 patients diagnosed with CCD, 1,011 (10%) underwent TCP, of whom 287 (28%) ultimately underwent PPM implantation (Fig 1). Among patients with evidence of CCD, 746 (8%) underwent PPM implantation without first undergoing TCP.
Overall, 11% of patients with Lyme carditis-associated CCD (N = 1,033) underwent PPM implantation. A majority of the patients (N = 833, 82%) who underwent TCP had a diagnosis of complete AV block (Fig 2). SA node dysfunction, first-degree AV block and second-degree AV block were present in 9%, 8%, and 15% of patients who underwent TCP implantation, respectively. Among patients who underwent PPM implantation, 604 (58%) were diagnosed with complete AV block. SA node dysfunction, first-degree AV block, and second-degree AV block were present in 29%, 6%, and 11% of patients, respectively.

Impact of Lyme carditis on cardiovascular adverse events and in-hospital mortality
The incidence of acute heart failure, cardiogenic shock, and death among patients hospitalized with Lyme disease was 0.9%, 0.1%, and 0.6%, respectively. Acute heart failure occurred in 2.4% of patients with Lyme carditis. This was approximately four-fold higher than the incidence of acute heart failure in patients without Lyme carditis (0.7%; P<0.001). In addition, the incidence of cardiogenic shock in patients with Lyme carditis was five-fold higher than that in patients without Lyme carditis (0.5% vs 0.

Predictors of pacemaker implantation in patients with Lyme carditis
The results of the multivariable logistic regression analysis to determine the predictors of TCP, PPM implantation, and in-hospital mortality in patients hospitalized with Lyme carditis are shown in

Discussion
This study utilized the largest all-payer inpatient database in the United States to assess cardiovascular outcomes in patients hospitalized with Lyme disease over the 12-year period between 2003-2014. During the study period, the number of patients hospitalized annually with Lyme disease more than doubled. Our data revealed that among hospitalized patients 11% exhibited evidence of Lyme carditis, a large majority of whom were diagnosed with CCD (Fig 1). Among patients with Lyme carditis-associated CCD, 11% underwent PPM implant during the index hospitalization. This represents 1% of all patients hospitalized with Lyme disease. The independent predictors of PPM implantation include increased age, presence of CHB, SA node dysfunction, and cardiac arrest. Estimates based on insurance claims data suggest that approximately 329,000 individuals in the U.S. were diagnosed with and treated for Lyme disease annually between 2005 and 2010. This estimate increased to 476,000 annually between 2010 and 2018 [19][20][21]. A similar increase in Lyme disease cases was reported in Canada [22]. This trend has been attributed to increased prevalence of Lyme disease. These findings likely explain the increasing numbers of hospitalizations for Lyme disease and Lyme carditis between 2003-2014 observed in this study. The increase in the prevalence of some comorbidities during the study period may be attributable a least in part to aging of the population.
Our study found that older age and male sex were associated with higher odds of developing Lyme carditis in patients with Lyme disease (Table 4) [19]. The association between age and Lyme carditis is likely driven by multiple factors. Although our study revealed that several age-related comorbidities (e.g., vascular disease and chronic pulmonary disease) are independent predictors of Lyme carditis, it is possible that other factors that we could not measure (e.g., pre-existing, subclinical cardiac conduction system disease) contributed as well. Several explanations for sex-related differences have been proposed, including differences in healthseeking behavior and differential susceptibility to non-cutaneous forms of Lyme disease [10,23].
In this study, approximately 10% of patients hospitalized with Lyme disease and CCD underwent TCP. The rate of TCP implant was highest in patients with CHB (37%). More than 1 in 4 patients (28%) in our study who underwent TCP subsequently received PPM implants. Of the patients in this study with Lyme carditis who were diagnosed with CCD, 7.7% received PPM implant without first undergoing TCP. A majority of these patients were diagnosed with advanced CCD.
Guidelines for antibiotic treatment of Lyme disease are well-validated, however, data relevant to the management of conduction system disease in Lyme carditis are limited to case reports and small single-center studies [6,8,10]. Consequently, there are no formal, evidencebased guidelines for pacemaker therapy in patients with Lyme disease who present with CCD.
Recent guidelines for pacemaker therapy in patients with CCD state that PPM implant is contraindicated in the context of acute Lyme carditis (Class III indication) as the AV block is expected to resolve without recurrence [37]. Although a large majority of cases of Lyme carditis resolve within two weeks of antibiotic treatment, CCD may persist for up to 6 weeks and beyond. Some reports have shown that Lyme carditis can cause persistent AV block even with adequate antimicrobial therapy [2,4,30,[38][39][40][41]. This is one possible explanation for the range of rates reported for PPM implantation in patients with Lyme carditis: 10% in our study and 4-13% in other published studies [10,22,41].
The use of temporary transvenous pacemakers is usually limited to few days, and requires immobilization and continuous telemetry monitoring, with risk for dislodgement, infection or cardiac perforation [42]. The use of temporary-permanent pacemakers (TPP), has been proposed as a strategy to reduce the rate of PPM implant rate in cases of Lyme carditis where cardiac conduction system recovery is prolonged [22,[43][44][45]. However, there are only a few published reports of TPP use in patients with Lyme carditis, and further investigation is required [42,43].
It is possible that longitudinal monitoring of patients with Lyme disease and persistent conduction system abnormalities may reduce the rate of PPM implantation. The "Suspicious index in Lyme carditis" (SILC) score, which was described after the start of the time period included in our study, uses key risk factors for Lyme carditis to establish Lyme disease as the etiology of CCD [41]. Application of this score in patients with CCD may help prompt the initiation of antibiotic therapy and reduce the number of patients who undergo PPM implantation. Retrospective application of the SILC score showed that 55% of patients who underwent PPM implantation due to Lyme carditis-related AV block eventually exhibited resolution of the underlying conduction disease [41]. However, the SILC score does not provide guidelines for the use of PPM implant in patients with CCD confirmed to be due to Lyme disease. Our study was not designed to assess the impact of a systematic approach such as the SILC score to the diagnosis and treatment of Lyme carditis on PPM utilization in this patient population [22].
Limitations of the NIS dataset prevented us from performing a comprehensive analysis of the appropriateness of PPM implant in patients with Lyme carditis. The NIS dataset does not include data regarding the number of patients who exhibited resolution of CCD during hospital admission. In addition, the NIS does not provide outcomes data after the index hospitalization. For those patients who underwent PPM implant, the NIS dataset does not provide information on the percentage of time that pacing was utilized after device implantation. Thus, the percentage of patients with Lyme carditis in whom pacemaker implantation might have been safely avoided cannot be assessed with this database.

Limitations
Recognized limitations of the NIS database include potential errors in coding of diagnoses. Not all listed diagnoses are associated with a time of onset and resolution, limiting the ability to determine the sequence of events in some cases. Each NIS entry corresponds to a single hospitalization and transfer between hospitals can produce more than one entry for an individual patient. This study included only hospitalized patients and the results cannot be generalized to the outpatient population. In addition, the NIS does not contain post-hospitalization data. Therefore, it was not possible to analyze the number of patients in whom Lyme carditis-related CCD resolved with treatment. The NIS database also does not include data on the utilization of cardiac pacing in those patients who underwent PPM implantation. Furthermore, post-hospitalization outcomes of patients with and without Lyme carditis could not be assessed with this database.