Article Text

Download PDFPDF

Original research
Characteristics of infective endocarditis in a cancer population
  1. Cullen Grable1,
  2. Syed W Yusuf1,
  3. Juhee Song2,
  4. George M Viola3,
  5. Owais Ulhaq1,
  6. Jose Banchs1,
  7. Corey T Jensen4,
  8. Harsh Goel1 and
  9. Saamir A Hassan1
  1. 1Department of Cardiology, Division of Internal Medicine, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
  2. 2Department of Biostatistics, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
  3. 3Department of Infectious Diseases, Infection Control and Employee Health, Division of Internal Medicine, The University of Texas MD Anderson Cancer Center, Houston, Texas, Houston, Texas, USA
  4. 4Department of Abdominal Imaging, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
  1. Correspondence to Dr Saamir A Hassan; SAHassan1{at}


Background Infective endocarditis (IE) is more common in patients with cancer as compared with the general population. Due to an immunocompromised state, the need for invasive procedures, hypercoagulability and the presence of indwelling catheters, patients with cancer are particularly predisposed to the development of IE.

Objectives Limited information exists about IE in patients with cancer. We aimed to evaluate the characteristics of patients with cancer and IE at our tertiary care centre, including a comparison of the microorganisms implicated and their association with mortality.

Methods A retrospective chart review of patients with cancer who had echocardiography for suspicion of endocarditis was conducted. A total of 56 patients with a confirmed diagnosis of cancer and endocarditis, based on the modified Duke criteria, were included in the study. Baseline demographics, risk factors for developing IE, echocardiography findings, microbiology and mortality data were analysed.

Results Following the findings of vegetations by echocardiography, the median survival time was 8.5 months. Staphylococcus aureus was the most common organism identified as causing endocarditis. The mitral and aortic valves were the most commonly involved sites of endocarditis. Patients with S. aureus endocarditis (SAE) had a significantly poorer survival when compared with patients without SAE (p=0.0217) over the 12-month period from diagnosis of endocarditis.

Conclusions Overall survival of patients with cancer and endocarditis is poor, with a worse outcome in patients with SAE.

  • endocarditis
  • infection
  • epidemiology

Data availability statement

All data relevant to the study are included in the article or uploaded as online supplemental information

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Key questions

What is already known about this subject?

  • Infective endocarditis (IE) occurs in more commonly in patients with cancer as compared with the general population due to a myriad of factors including an immunocompromised state, the need for invasive procedures, hypercoagulability and the presence of indwelling catheters.

What does this study add?

  • There is limited information about the characteristics of IE in patients with cancer. Our study shows that overall survival of patients with cancer and endocarditis is poor (median survival of 8.5 months), IE occurs most commonly on the mitral and aortic valves, and that Staphylococcus aureus endocarditis has worse outcome.

How might this impact on clinical practice?

  • Our study illustrates that there must be a high index of suspicion and a thorough investigation in patients with cancer with suspected IE. Given the poor prognosis of IE in patients with cancer, it is of the utmost importance that these patients are efficiently diagnosed and treated.


Endocarditis is a relatively uncommon but potentially fatal disease affecting 3–10 per 100 000 people, with a 1-year mortality rate of almost 30%.1–4 The demographics of patients infected with endocarditis have been changing. Until the 1980s, infective endocarditis (IE) was most prevalent in patients between 40 and 50 years of age and was mainly due to streptococcal organisms.1 5 6 IE is now most common in patients older than 70 years.7 Advances in therapy across many areas of medicine have led to the increasing use of indwelling devices such as prosthetic heart valves and long-term intravenous lines, which in turn have led to increased prevalence of staphylococcal bacteraemia, a precursor to endocarditis.8–10 This has also led to the emergence of Staphylococcus aureus as the most common organism in IE in some series, with >25% of cases being due to this organism.4

Due to an immunocompromised state, the need for invasive procedures, hypercoagulability, the presence of indwelling catheters and pre-existing valvular heart disease (VHD) (figure 1), patients with cancer are particularly predisposed to the development of IE.11–13 In addition, patients with cancer often develop recurrent local or systemic infections, which increases the possibility of IE. IE is common in patients with cancer, with a prevalence of about 18% in one study.14 In another prospective, multicentre study, 5.6% of patients with endocarditis had cancer.15 IE also has grave consequences in patients with cancer, as it is associated not only with increased mortality as compared with non-cancer patients but also cessation of chemotherapy and denial of more aggressive interventions.2 14 15 Limited information exists about the associated risk of IE with different types of cancer and the microorganisms commonly involved in IE in patients with cancer. A better understanding of predisposing factors may lead to improved identification and management of IE in this at-risk population of patients. Herein, we aimed to evaluate the characteristics of patients with cancer and IE at our tertiary care centre, including a comparison of the microorganisms implicated and their association with 12-month survival following the diagnosis of endocarditis.

Figure 1

Conditions in patients with cancer that are conducive to infective endocarditis. RBC, red blood cells.

Figure 3

Kaplan-Meier survival curves by presence of Staphylococcus aureus endocarditis. The curves were censored at 12 months after endocarditis diagnosis.


Patient selection and data collection

We conducted a retrospective chart review of 1839 patients with cancer who had echocardiography for suspicion of endocarditis from May 2001 to December 2006 and from July 2015 to December 2018 at The University of Texas MD Anderson Cancer Center (figure 2). Records in between these dates were not accessible due to electronic health record transition made by our institution during the study period. Among these patients, we identified 56 with a confirmed diagnosis of cancer as well as endocarditis, based on the modified Duke criteria.16 Data were extracted from the patient charts and analysed to confirm the diagnosis of endocarditis as per the modified Duke criteria. Data collected included baseline demographics (including age at diagnosis of cancer and IE, sex and race/ethnicity), findings of echocardiography (including transthoracic and transoesophageal studies), type of cancer, use of chemotherapeutics, use of steroids and mortality data. We also analysed the traditional risk factors for IE, including VHD, a pacemaker, a central line, dental surgery, recent invasive procedures, recent surgery, diabetes mellitus, HIV, heart failure and chronic kidney disease. Lastly, we evaluated the presence of modified Duke major and minor criteria for endocarditis, baseline serum chemistry data, septic emboli based on imaging studies, as well as blood culture results.

Figure 2

Flow diagram for patient inclusion.

Patient and public involvement

Patients and the public were not involved in the design, conduct, or reporting of this study as it was done retrospectively.


The blood cultures were processed by the Microbiology Laboratory at MD Anderson Cancer Center at the time of IE diagnosis. After the samples were processed, standard techniques were employed to determine which pathogen was responsible, as well as antimicrobial sensitivities.17 18


The modified Duke criteria were used to identify patients with endocarditis.16 19 Blood cultures that grew S. aureus and other typical organisms associated with endocarditis were considered clinically significant as the likely causative organism for IE. Patients with platelet counts less than 50 000/µL were considered thrombocytopaenic. Neutropaenia was defined to be absolute neutrophil count (ANC) of less than 500 cells/µL. Active cancer was defined as any cancer that was diagnosed or being managed by cancer therapeutics at the time of diagnosis of IE. Culture-negative IE was defined as endocarditis without microbiological aetiology following inoculation of samples in a standard blood-culture system with negative cultures after 7 days of incubation and subculturing. Predisposing valvular heart disease (VHD) was defined as patients at risk of VHD (stage A) or with progressive VHD (stage B) to severe asymptomatic (stage C) and symptomatic (stage D) VHD as per the American College of Cardiology/American Heart Association (AHA) societal guidelines, including patients with patients history of rheumatic heart disease, mitral valve stenosis, mitral regurgitation, bicuspid aortic valve, aortic stenosis, aortic regurgitation, prosthetic heart valve, history of IE, history of congenital heart disease or abnormal thickening of heart valves with greater than mild regurgitation.

Statistical analysis

Patient demographic and other characteristics were summarised by S. aureus status (presence or absence) using descriptive statistics. Survival time over 12 months (12-month survival) from diagnosis of endocarditis was calculated from first transthoracic or transoesophageal study to death or last follow-up occurred within 12 months. For those who survived longer than 12 months were censored at 12 months. Univariate and multivariate Cox proportional hazards regression models were used to identify factors associated with 12-month survival. Multivariate Cox models included factors with a significant p value based on univariate Cox models. The proportional hazards assumption was checked by the interaction between each covariate and the logarithm of survival time. When the proportional hazards assumption was violated, the time–covariate interaction was included in the multivariate model. A p<0.05 indicated statistical significance. SAS V.9.4 software (SAS Institute) was used for statistical analysis.


Patient characteristics

We included 56 patients with a concurrent diagnosis of cancer and IE (table 1). The mean patient age was 54 years at diagnosis of cancer and 60 years at diagnosis of endocarditis. Most patients were white (45%) and male (55%). Thirty-six (64%) had active cancer. More than half of the patients had an underlying solid cancer (68%), while 18 patients (32%) had a haematological malignancy. The most common comorbidities among the cohort were diabetes (77%), chronic kidney disease (18%), and congestive heart failure (13%). Thirty-nine patients (70%) had culture-positive endocarditis and 17 (30%) had culture-negative endocarditis. S. aureus was the most common organism isolated in this cohort (20 patients), while 10 patients had Enterococcus species, 7 had Streptococcus viridans, 1 had a HACEK organism (Klebsiella) and 1 had Streptococcus gallolyticus. Among patients with S. aureus endocarditis (SAE), 11 patients were infected with methicillin-resistant S. aureus (MRSA). The most commonly involved site was the mitral valve, in 23 (41%) patients, followed by the aortic valve in 20 (36%) patients and the tricuspid valve in 9 (16%) patients. Two patients (4%) had pacemaker lead-associated endocarditis. Of the 23 patients with mitral valve endocarditis, seven (30%) had evidence of systemic embolisation, and four patients had vegetations ≥10 mm (17%). Three of four patients with vegetations ≥10 mm had systemic emboli. Of the 20 patients with aortic valve endocarditis, four (20%) had systemic emboli, one had large vegetations ≥10 mm and one had an abscess. Of the nine patients with tricuspid valve endocarditis, two (22%) had evidence of pulmonary infarcts, suggesting small pulmonary emboli. Twenty of the 56 patients (36%) had a central venous catheter at the time of IE diagnosis. Twenty-one patients had an indwelling foley catheter (38%), with five patients having a nephrostomy tube (9%) and one patient having a urostomy with an ileal conduit (2%).

Table 1

Patient characteristics

Table 2 compares the characteristics of patients with SAE and those without S. aureus. Patients with SAE had a significantly higher white cell count (WCC) (14.68 K/µL vs 11.29 K/µL in those without SAE; p=0.044) and had a significantly lower percentage of predisposing VHD (0% vs 22% in those without SAE; p=0.04). Otherwise, there were no significant differences between the two groups in age, sex, race/ethnicity, cancer status, comorbidities, indwelling catheter or location of IE involvement.

Table 2

Comparison of patient characteristics with and without Staphylococcus aureus endocarditis

IE outcomes

After diagnosis of IE by echocardiography, the median survival time for our cohort was 8.5 months. In univariate analysis, black race was associated with significantly poorer 12-month survival when compared with all other races/ethnicities (HR 2.320; 95% CI 1.114 to 4.833; p=0.0246). Over the 12 months after diagnosis of IE, patients with active cancer had significantly poorer survival when compared with patients in remission (HR 2.497; 95% CI 1.062 to 5.868; p=0.0358).

Furthermore, increasing WCC, blood urea nitrogen, alanine aminotransferase (ALT), aspartate aminotransferase (AST), WCC <4000 cells/µL and >11 000 cells/µL, ALT ≥41 U/L, AST ≥40 U/L and bilirubin ≥1.2 mg/dL (above-normal institutional laboratory values) were significantly associated with increased risk of death within 12 months (table 3). A total of seven patients had cardiac valve surgery for endocarditis; six of them had SAE. Surgical treatment was not significantly associated with increased risk of death (HR 0.671; 95% CI 0.086 to 5.242; p=0.7036).

Table 3

Univariate Cox regression analysis: variables that were significantly associated with survival within 12 months

Although all patients were followed by an infectious diseases specialist and treated accordingly to AHA guidelines,20 patients with SAE had significantly worse 12-month survival when compared with those without SAE (HR 2.305; 95% CI 1.107 to 4.802; p=0.0257). The use of steroids was also associated with an increased risk of death within 12 months (HR 8.151; 95% CI 3.059 to 21.719; p<0.0001).

Multivariate Cox regression analyses revealed that the presence of SAE (HR 2.52; 95% CI 1.193 to 5.323; p=0.0015) and active cancer status (HR 2.967; 95% CI 1.253 to 7.025; p=0.013) and black race (HR 2.209; 95% CI 1.038 to 4.701, p=0.0396) remained significant risk factors for 12-month survival (table 4, model 1). The proportional hazards assumption was violated for SAE. Therefore, HR for SAE was interpreted as the weighted average over follow-up in the model without time–covariate interaction (table 4, model 1). In a second model (table 4, model 2), we considered time and SAE interaction (allowing two different HRs for two-time intervals, ≤4 months and >4 months of follow-up). Presence of SAE was not significantly associated with overall survival during the first 4 months but was significantly associated with overall survival after 4 months (HR 17.316; 95% CI 1.993 to 150.44; p=0.0097). Active cancer status (HR 2.892; 95% CI 1.217 to 6.875; p=0.0162) and black race (HR 2.140; 95% CI 1.008 to 4.453; p=0.0476) was significantly associated with 12-month survival. The Kaplan-Meier survival curve comparing patients with SAE versus non-SAE also reflects significantly poorer survival with SAE (p=0.0217) over the 12-month period from diagnosis of endocarditis (figure 3).

Table 4

Multivariate Cox regression model


Patients with cancer with IE have poor prognosis,15 and among such patients, those with SAE may in fact have the worst mortality after IE diagnosis. Our study showed a statistically significant association between mortality and SAE compared with non-SAE in patients with cancer. In fact, after 4 months patients with SAE were 20 times more likely to die compared with non-SAE patients.

Few studies have focused on IE in patients with cancer, and there is a relative paucity of data about the most prominent types of cancer associated with IE, which valves/sites are involved, and rates due to venous lines or cardiac devices. One study of 161 patients with cancer and IE showed that the majority of patients with solid tumours had intestinal cancers (42%) and the majority of those with haematological malignancies had lymphoma (42%).15 Studies have also showed that the vast majority of vegetations involve the mitral and aortic valves in patients with and without cancer.3 15 The source of the IE was documented to be associated with a central venous catheter in 23% of patients in a previous study of patients with cancer and IE, and up to 10% of the vegetations were associated with pacemaker leads.15 Our study showed that the mitral valve was most commonly involved, followed closely by the aortic valve, and 4% of patients had pacemaker-lead endocarditis.

The typical causative organism for non-cancer patients with IE has been found to be S. aureus.3 21 22 Some studies report Streptococcus spp most commonly, followed by S. aureus.15 23 One study of patients with cancer with IE showed equal prevalence of S. aureus and Streptococcus spp, with each comprising 23.6% of cases.15 Others have shown S. aureus to be the most common causative organism for IE in patients with cancer.24 Our study found that 20 (51%) of the 39 patients with culture-positive endocarditis had presence of S. aureus; this was 36% of all the patients (20 of 56).

SAE has been associated with a significant increase in in-hospital death in non-cancer patients and significantly increased 30-day mortality in surgical patients.3 22 23 Fernández-Cruz et al showed SAE to be a risk factor for 30-day mortality in patients with cancer with IE.15 Our study builds on this association of SAE and poor outcomes by showing significantly worse 12-month survival when compared with culture-negative IE and presence of other organisms. S. aureus is more problematic due to multiple factors, including its ability to adhere to and multiply on cardiac cells, form biofilms preventing eradication and rapidly develop resistance.22 25 26

Compared with the general population, patients with cancer are more likely to die from IE, but the mechanism is not completely known.27 Patients with cancer are more frequently neutropaenic, may be immunocompromised because of chemotherapy or the malignancy itself, and frequently have indwelling catheters that may serve as a source of infection.28 Specifically, cancer may predispose to IE by facilitating not only bacterial translocation into the bloodstream but also localisation to the endocardium through existence of non-bacterial thrombotic vegetations that may be preexisting due to the hypercoagulable state associated with cancer.27 29 Among patients with solid malignancies, patients with intestinal tumours have been found to be most likely to develop IE with Streptococcus gallolyticus (formerly known as S. bovis).15 This may be due to the tumour facilitating intestinal bacterial translocation to the bloodstream. In haematological malignancies, lymphoma has been found to be the most prevalent cancer associated with IE.15 Particularly, S. aureus IE has been associated with cutaneous portals of entry.30 This, combined with the frequency of indwelling catheters, may be the reason for patients with cancer with IE most frequently grow S. aureus.

Thirty per cent of the patients in our study had blood culture–negative endocarditis. Our group previously reported a 42% frequency of culture-negative endocarditis in a retrospective series of patients with cancer.6 This culture-negative rate may be due to a variety of factors including early administration of antibiotics,3 non-infectious aetiologies of the vegetations, and lack of use of PCR techniques to identify organisms.31 32 This yield may be increased by using reverse transcription-PCR in addition to culture, valvular biopsies, serological analysis and evaluation for non-infectious causes, including autoimmune and thrombotic.31 32 Some of these patients may have had non-bacterial thrombotic endocarditis, which can be found in patients ith cancer; previous studies have shown that this form of endocarditis can be associated with an embolic phenomenon in solid-tumour cancers such as lung, pancreatic and gastric cancer.33 34 Although culture negative patients in our study were classified as definitive IE by the modified Duke’s Criteria, its diagnostic specificity is estimated to only be 74% based on the results of a recent cohort study.35 This is further complicated by the fact that no single laboratory, imaging, nor clinical criteria can definitively diagnose marantic endocarditis. Therefore, distinguishing true culture negative IE from marantic endocarditis is challenging and often requires a high degree of clinical suspicion, with definitive distinction often made during autopsy.36 The use of fluoro-18-fluorodeoxyglucose positron emission tomography/CT and radiolabeled WCC scintigraphy has been explored to identify underlying infectious valvular foci, although the diagnostic utility of these imaging techniques appears to be greatest with suspected prosthetic valve and cardiac implantable electronic device (CIED) endocarditis.37

Endocarditis frequently is believed to originate from cutaneous portals of entry into the bloodstream.30 The most common healthcare-associated causes of cutaneous portals are vascular access and implantable cardiac devices and prosthetic valves.3 30 For example, within the first year of valve replacement, infections with coagulase negative staphylococci, S. aureus, HACEK organisms and fungi are much more common, whereas Streptococcus spp and Enterococcus infections occur less frequently.38 This microbiology reflects the nosocomial nature of prosthetic valve endocarditis acquired perioperatively. Of note, however, prosthetic valve endocarditis after the first year of valve replacement tends to be community acquired, as reflected by increased incidence of Streptococcus spp and Enterococcus infections.38 Antibiotic prophylaxis for staphylococci prior to implantation of a CIED is a class 1 recommendation by the AHA and was shown to be beneficial in a randomised clinical trial.39 40 However, antibiotic prophylaxis in minor invasive procedures has been challenged and remains under debate.26 It has been suggested that antibiotic prophylaxis be used in invasive procedures for those with congenital heart defects because of increased risk of infection.23 Perhaps prophylaxis with antibiotics should be considered before procedures in patients with cancer due to the myriad factors discussed above. However, both the Society of Interventional Radiology and American Society of Clinical Oncology (ASCO) recommend against routine antibiotic prophylaxis in the general cancer population in the setting of placement of a totally implantable vascular access device (ie, a chemotherapy port).41 42 Instead, they recommend ensuring a clean procedure is done with the use of central venous catheter clinical care bundles. These guidelines were based on prior surgical literature and retrospective studies demonstrating lack of clinical benefit with routine antibiotic prophylaxis and low baseline rate of device-associated infections. Further supporting evidence came from two randomised controlled studies that demonstrated no difference in the rate of device-associated infections between antibiotic-prophylaxis and placebo groups. However, these studies were limited by their small sample sizes, low incidence of infectious complications, exclusion of liquid tumours in one study, and exclusion of patients with neutropaenia in both.43 44 A meta-analysis of these studies combined with two prior retrospective studies recapitulated the lack of benefit of antibiotic prophylaxis.45–47 Interestingly, a recent clinical trial found that universal MRSA decolonisation in intensive care unit patients resulted in a greater reduction in all-cause bloodstream infections and MRSA clinical isolates compared with targeted decolonisation (MRSA screening followed by decolonisation) and isolation measures.48 However, although targeted MRSA decolonisation has been studied extensively in surgical populations, concerns for cost-effectiveness and the development of multidrug-resistant MRSA remain with widespread adoption of this approach.49

There is conflicting evidence regarding the risk of device-related infectious complications and neutropaenia at the time of device placement, but a 2019 retrospective single-centre cohort study addressed this question and found a significantly greater rate of infection-related port removal within 30 days of device placement in neutropaenic patients.50 However, the study was limited by its retrospective nature, a higher rate of liquid tumours in the neutropaenic cohort and lack of documentation of pre-procedure antibiotic prophylaxis. There are currently no specific recommendations by the Infectious Diseases Society of America or ASCO on the use of antibiotic prophylaxis in the neutropaenic and liquid tumour patient populations. Future studies are needed to help further assess which patients, if any, may benefit from antibiotic prophylaxis prior to invasive procedures.


Among patients with cancer treated at our tertiary care centre, the development of endocarditis portended poor survival, and the presence of SAE or active cancer further increased mortality. This study further contributes to the limited knowledge about endocarditis, particularly SAE, in the cancer population. Given these findings, we emphasise the need for further research in prevention and treatment of this life-altering complication.

Data availability statement

All data relevant to the study are included in the article or uploaded as online supplemental information

Ethics statements

Patient consent for publication

Ethics approval

The study was approved by the institutional review board, and the requirement for patient consent was waived.


We thank Sunita Patterson, Research Medical Library, for editing this article.



  • Correction notice Since the publication of this article, initials have been added to the following author names: Saamir A Hassan and Syed W Yusuf.

  • Contributors CG and SAH contributed to the planning, conduct, writing and reporting of the report. SWY, JS, GMV, OU, JB, CTJ and SAH contributed to the planning, conduct and reporting of the work. SAH will serve as the guarantor of the work.

  • Funding Supported in part by the National Institutes of Health/National Cancer Institute through award number P30CA016672 (used the Clinical Trials Office and Biostatistics Resource Group).

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles