Cardiac valve replacement for infective endocarditis in patients with end stage renal disease on hemodialysis — A single institution experience

Introduction

Complications from vascular access are the leading cause of morbidity in the hemodialysis (HD) population while infection is the second leading cause of death in this population. Incidence of infections affecting HD access sites ranges from 0.56 to 5% per year for autogenous access and 4 to 20% per year for prosthetic grafts.^{1, 2} While the overall incidence of bacteremia ranges from 0.7 to 1.2 episodes per 100 patient months, use of tunneled catheters (TCs) is associated with a greater risk when compared to arteriovenous fistulae (AVF) and prosthetic arteriovenous grafts (AVGs) accounting for approximately 3.9–16.7 episodes per 100 patient months.^3,4

Infective endocarditis (IE) has been recognized as a serious complication of chronic HD since the 1960s occurring in 2–12% of patients.^1,3–6 Although the overall incidence of IE is low (0.0045%), it is significantly higher than in non-dialysis-dependent patients with a relative risk of 16.9.^{1,7, 8} This phenomenon is likely secondary to a few contributing factors including bacterial burden from repetitive vascular access with microbial seeding, underlying valvular heart disease, and immunocompromised state. ¹ Thirty-day mortality rates are also higher in HD patients with IE compared to non-dialysis-dependent patients. ⁷ As expected, many studies show that patients with synthetic HD access (TC or AVG) are more likely to develop IE than those with autogenous access (AVF); however, few reports show that more IE cases develop in those with autogenous access.

Treatment of IE involves appropriate long-term antibiotics; however, depending on the virulence of the microorganism and the extent of involvement of surrounding tissues, surgical treatment may become indispensable to eradicate the infection. Although wrought with high morbidity and mortality rates due to the severely ill patient population, surgical treatment has become the treatment of choice in 20–50% of cases of complicated IE.

Several previous studies have documented outcomes in patients with end stage renal disease (ESRD) and IE, ESRD requiring valve replacement (VR), and IE requiring VR.^{5,9, 10} However, there are only a few reported series in patients with both ESRD and IE requiring VR. ¹ Therefore, our objective was to review outcomes in HD-dependent ESRD patients requiring VR for IE.

Methods

After approval was obtained from the institutional review board, a retrospective chart review was conducted using Apollo Clinical Database to identify all patients with ESRD on HD who underwent VR for IE at Montefiore Medical Center within a six-year period (January 2009–December 2014). Inclusion criteria included a diagnosis of IE and ESRD on HD. Exclusion criteria included patients with ESRD on peritoneal dialysis and acute renal failure requiring HD.

Using Montefiore Medical Center’s Clinical Looking Glass Database and Electronic Medical Records, relevant clinical information including demographics (age at date of VR, gender, and race/ethnicity), risk factors (smoking status, hemoglobin A1c (HbA1c), and body mass index (BMI)), comorbidities (hypertension (HTN), diabetes mellitus (DM), hyperlipidemia (HLD), cerebrovascular accident (CVA), peripheral arterial disease, and human immunodeficiency virus (HIV)), cardiac disease, coronary artery disease (CAD), congestive heart failure (CHF), atrial fibrillation, prior coronary artery bypass graft, prior valvular disease, and prior valve repair/replacement), valve characteristics (valve involved and causative organisms), and type of HD access (AVF, AVG, or TC) was collected. The clinical outcome recorded was 30-day mortality for the total study population and specific subsets (IE, ESRD on HD, and IE and ESRD on HD). Descriptive statistics were used to summarize demographic and clinical characteristics.

Results

A total of 1497 patients underwent VR within the six-year period. Of these, 167 patients (11.2%) had IE and 119 patients (7.9%) had ESRD on HD. Of these, 33 patients (2.2%) had both IE and ESRD on HD (Figure 1). The demographic and clinical characteristics of these 33 patients were further investigated.

OPEN IN VIEWER

Figure 1.

Study population. ESRD: end stage renal disease; HD: hemodialysis; IE: infective endocarditis; VR: valve replacement.

The mean age at surgery was 55 years; 19 (57.6%) of patients were male. The majority of the patients were African American (54.5%) (Table 1). With respect to risk factors for cardiac disease, most patients had never smoked (63.6%). The mean HbA1c was 5.7% and BMI was 27.4 kg/m² (Table 2).

OPEN IN VIEWER

Table 1.

Demographic characteristics.

	IE + HD (n = 33)
Age at surgery (years)
Mean (range)	55 (30–77)
Gender
Male	19 (57.6%)
Female	14 (42.4%)
Race
White/Caucasian	1 (3.0%)
African American	18 (54.5%)
Asian	1 (3.0%)
Multiracial/Other	10 (30.3%)

HD: hemodialysis; IE: infective endocarditis.

OPEN IN VIEWER

Table 2.

Risk factors.

	IE + HD (n = 33)
HbA1c (%)
Mean (range)	5.7 (3.9–12.2)
BMI (kg/m2)
Mean (range)	27.4 (15.4–56.6)
Smoking status
Current	4 (12.1%)
Former	8 (24.2%)
Never	21 (63.6%)

BMI: body mass index; HbA1c: hemoglobin A1c; HD: hemodialysis; IE: infective endocarditis.

All but one patient had HTN (97%). The Majority of patients had DM and HLD (57.6%), and one third of the patients (33.3%) had prior CVA. Only one patient (3%) had HIV (Table 2). Most of the patients had prior cardiac disease. CAD was prevalent in 16/33 (48.5%) of patients, while 10/33 (30.3%) had CHF. Nine of the 33 patients (27.3%) had prior valvular disease, and of these, three (33.3%) had prior valve repair/replacement (Table 3). Two patients (6.1%) were identified as IV drug users. Vascular access was documented as the likely source of infection in 16 patients (48.5%) with 22 patients (66.6%) having positive blood cultures. Only five (15.2%) had surgical interventions to address local infections; however, 13 patients (39.4%) underwent catheter changes to address their infection. Twenty-three patients (69.7%) were treated with IV antibiotics for an average of three weeks.

OPEN IN VIEWER

Table 3.

Clinical characteristics.

	IE + HD (n = 33)
Comorbidities
HTN	32 (97.0%)
DM	19 (57.6%)
HLD	19 (57.6%)
CVA	11 (33.3%)
PAD	8 (24.2%)
HIV	1 (3.0%)
Cardiac disease
CAD	16 (48.5%)
CHF	10 (30.3%)
Afib	7 (21.2%)
Prior CABG	11 (33.3%)
Prior valve disease	9 (27.3%)
Prior valve repair	3 (9.1%)

CABG: coronary artery bypass graft; CAD: coronary artery disease; CHF: congestive heart failure; CVA: cerebrovascular accident; DM: diabetes mellitus; HD: hemodialysis; HIV: human immunodeficiency virus; HLD: hyperlipidemia; HTN: hypertension; IE: infective endocarditis; PAD: peripheral arterial disease.

The majority of the patients 17/33 (51.5%) were being dialyzed via TC, while 12/33 (36.4%) were being dialyzed via AVF and 4/33 (12.1%) via AVG (Table 5). Seventeen patients (51.5%) kept the same HD access postoperatively. The average length of HD prior to developing endocarditis was 13.6 months.

OPEN IN VIEWER

Table 4.

Valve characteristics.

	IE + HD (n = 33)
Valve involved
Aortic	13 (39.4%)
Mitral	19 (57.6%)
Aortic and mitral	1 (3.0%)
Pathogenic organism
MRSA	6 (18.2%)
MSSA	11 (33.3%)
Coagulase-negative Staph	7 (21.2%)
Enterococcus	1 (3.0%)
Gram negative	1 (3.0%)
Candida	1 (3.0%)
Other	1 (3.0%)
Polymicrobial	2 (6.1%)
Negative Bcx	1 (3.0%)
N/A	2 (6.1%)

HD: hemodialysis; IE: infective endocarditis; MRSA: Methicillin-resistant Staphylococcus aureus; MSSA: Methicillin-sensitive Staphylococcus aureus.

The mitral valve was affected in the majority of the patients (57.6%), followed by the aortic valve (39.4%) (Table 4). Staphylococcus species accounted for the majority (73%) of the pathogens, and of these, Methicillin-sensitive Staphylococcus aureus (MSSA) was the most prevalent accounting for 33.3% of patients (Table 4). Indications for surgery included valvular insufficiency in 30 (90.1%) as well as fungal endocarditis and septic embolization.

OPEN IN VIEWER

Table 5.

Type of HD access.

	IE + HD (n = 33)
AVF	12 (36.4%)
AVG	4 (12.1%)
TC	17 (51.5%)

AVF: arteriovenous fistula; AVG: arteriovenous graft; HD: hemodialysis; IE: infective endocarditis.

Overall 30-day mortality for VR was 5.0% (75/1497). Mortality for patients with IE was 7.2% (12/167) and for patients with ESRD on HD was 10.1% (12/119). Mortality occurred in 2 of 12 patients with AVF (16.7%), 1 of 4 patients with AVG (25%), and 2 of 17 patients with TC (11.8%) for an overall mortality of 15.2% (5/33) for patients with IE and ESRD on HD (Table 6).

OPEN IN VIEWER

Table 6.

Thirty-day mortality.

VR	75/1497 (5%)
IE	12/167 (7.2%)
ESRD	12/119 (10.1%)
IE + ESRD	5/33 (15.2%)
AVF	2/12 (16.7%)
AVG	1/4 (25.0%)
TC	2/17 (11.8%)
IE − ESRD	7/134 (5.2%)
ESRD − IE	7/86 (8.1%)
IE + ESRD	5/33 (15.2%)

AVF: arteriovenous fistula; AVG: arteriovenous graft; ESRD: end stage renal disease; IE: infective endocarditis; TC: tunneled catheter; VR: valve replacement; ESRD: end stage renal disease; IE: infective endocarditis.

Thirty-day mortality rate for patients with IE without ESRD was 5.2%, whereas for those with ESRD without IE was 8.1%. However, for patients with IE and ESRD, the mortality rate was 15.2% (Table 6). Causes of death included sepsis, cardiac arrest, heart failure, lower GI bleed, and pneumonia.

Discussion

Despite intensive programs to curb their use, 81% of ESRD patients initiate HD with a catheter while only 26% have an autogenous or prosthetic AV access already in place. ² In our study population, the majority of patients (approximately 64%) had nonautogenous AV access (AVG or TC) only 36.4% had AVFs. We expected that almost all of the patients with ESRD on HD who developed IE would likely have nonautogenous access as it poses a significant risk for bacteremia, sepsis, and IE likely due to migration of skin flora during cannulation of the access or contamination of catheter or graft lumens. ¹

Previous studies have isolated Staphylococcus species in the majority of ESRD patients with IE.^{1, 2} Similarly, our study confirms that Staphylococcus species were indeed the most prevalent (73%) with MSSA being the most frequent among these. Of the coagulase-negative Staphylococci species, most were Staphylococcus lugdunensis, which has a high potential for valve destruction. These organisms are usually skin flora, which lends further credence to the fact that the IE was more likely related to HD access.

The mitral valve has been reported to be the most commonly affected valve, which has been confirmed in this study (57.6%). This is likely explained by mitral annular calcification and valvular thickening frequently seen in ESRD patients, which via alterations of laminar flow, leads to an increased susceptibility for IE. ¹ A significant proportion of our study population also had prior valvular disease (33.3%).

Two-point-two percent of the VR were performed for IE in ESRD pts. The overall operative 30-day mortality in previous studies for IE have ranged from 6 to 26% for native and 8 to 67% for prosthetic valve.^{1, 10} Our 30-day mortality rate (7.2%) for patients with VR and IE was comparable.

Studies show that comparing patients undergoing VR with patients treated medically, the mortality almost doubles (40% versus 73%). ¹ This is likely due to the fact that patients considered for VR surgery either were not candidates for medical management to begin with or had undergone conservative treatment with antibiotics that failed, suggesting more severe disease in the patients considered for surgery. ¹ There is controversy regarding whether accepted indications for VR in the general population are applicable to patients with ESRD. ¹

Despite mortality rates between 42 and 73% reported in the literature in patients with ESRD and IE who have had VR, our mortality rate was 15.2%. ¹ The 30-day mortality rates for different subsets were also calculated. Mortality rate was higher for patients with both IE and ESRD than either alone (Table 6). There was no difference in mortality rates among those with AVF, AVG, or TC.

Limitations to this study include the relatively small sample size analyzed and the retrospective nature of the report. This can primarily be attributed to the low incidence of IE. Additional analysis can focus on mortality of patients with IE. It should also be pointed out that one of the patients in our study had HIV, potentially making them more susceptible to infection, sepsis, and other complications. In the future, prospective studies could be conducted to better characterize this population. Our ultimate aim is to discover the incidence of catheter/access related bacteremia causing endocarditis and of the patients affected, how many went on to require surgery. This further lends credence to the slogan “catheter last.” These results bolster the national focus on the Fistula First Breaktrough Initiative and Catheters Last.

Conclusion

IE remains a significant problem in patients with ESRD on HD, particularly when TCs are utilized for HD access. Although appropriate algorithms have been developed to minimize long term use of TCs, bacteremia remains a significant problem. We reviewed our institutional experience and the medical literature to determine outcomes in HD-dependent patients with IE requiring VR. Despite mortality rates between 42 and 73% reported in the literature, our mortality rate was 15.2%. ¹ Care of these critically ill patients must emphasize early diagnosis and aggressive management to optimize outcomes. To our knowledge, this is one of the largest retrospective reviews of patients with ESRD and IE requiring VR.