By Richard R. Watkins, MD, MS, FACP, FIDSA

Associate Professor of Internal Medicine, Northeast Ohio Medical University; Division of Infectious Diseases, Cleveland Clinic Akron General, Akron, OH

Dr. Watkins reports that he has received research support from Allergan.

SYNOPSIS: Using large databases from New York and California, investigators found the overall incidence of infective endocarditis remained stable between 1998 and 2013, and 90-day mortality declined. Changes were noted in pathogen etiology and patient characteristics over time.

SOURCE: Toyoda N, Chikwe J, Itagaki S, et al. Trends in infective endocarditis in California and New York state, 1998-2013. JAMA 2017;317:1652-1660.

Infective endocarditis (IE) remains an uncommon yet serious illness. In 2007, a major change occurred in the IE prophylaxis guidelines, with the recommendation of fewer indications for prophylaxis. In light of this change, Toyoda et al used large databases to examine trends in the epidemiology and outcomes of IE between 1998 and 2013.

Patients for the study were identified using ICD-9 codes in statewide databases from New York and California. These databases included information on every hospital discharge, ambulatory surgery, and emergency room visit in their respective state. IE was characterized as native valve, prosthetic valve, cardiac device-related, or drug abuse-associated. Primary and secondary diagnostic codes were used to identify causal microorganisms, including Staphylococcus aureus (methicillin-resistant [MRSA] and methicillin-susceptible [MSSA]), other Staphylococcus species, Streptococcus species, gram-negative organisms, fungi, and unknown, which included both culture-negative cases and those without a code.

During the study period, 75,829 cases of first-episode IE were identified, 56% in California and 43% in New York. The crude annual incidence increased from 7.6 to 9.3 cases per 100,000 persons. However, after adjustment for age, sex, and race, there was no significant increase in IE over time (range, 7.6 to 7.8 cases per 100,000 annually), and 90-day mortality decreased annually by approximately 2%. During the latter part of the study, those diagnosed with IE tended to be older, more likely to be male, and more likely to have chronic obstructive pulmonary disease, cancer, or liver disease. Drug use-associated IE increased over the study period by 0.9% annually (95% confidence interval [CI], 0.4-1.3). There was a substantial increase in hemodialysis patients diagnosed with IE, from 14.9% to 17.9% and representing 35.0% of healthcare-associated cases of IE between 2010 and 2013.The proportion of patients with a history of valve surgery increased from 12.8% to 15.2%, and the proportion with implantable cardiac devices increased from 8.8% to 15.6%.

Overall, these trends resulted in a decreased proportion of patients with native-valve IE at the end of the study (74.5% to 68.4%) and an increased proportion with prosthetic-valve IE (12.0% to 13.8%) and device-related IE (1.3% to 4.1%). The proportion of healthcare-associated IE increased from 49.8% in 1998 to 51.2% by 2013. The standard incidence of S. aureus IE increased during the study period from 2.1 (95% CI, 2.0-2.2) to 2.7 (95% CI, 2.6-2.9) cases per 100,000 people annually, with MRSA increasing from 0.23 (95% CI, 0.19-0.27) to 1.13 (95% CI, 1.05-1.22). The incidence of oral streptococcal IE decreased from 0.84 (95% CI, 0.76-0.92) to 0.73 (95% CI, 0.67-0.80) cases per 100,000 people annually and there was not an increase after the prophylaxis guidelines were changed. More patients underwent cardiac surgery during or within 30 days of their index admission over the study period (10.6% to 13.3%). Finally, healthcare-associated IE was associated with greater mortality compared to community-onset IE (adjusted hazard ratio [aHR], 1.52; 95% CI, 1.48-1.56), and compared to streptococcal IE, mortality was greater with gram-negative IE (aHR, 1.22; 95% CI, 1.16-1.28), staphylococcal IE (aHR, 1.38; 95% CI, 1.34-1.42), and highest with fungal IE (aHR, 1.84; 95% CI, 1.72-1.99).


This is an important study for several reasons: the large number of patients that were included, the timeframe studied, which ranged from before and after the guidelines changed for IE prophylaxis, and the findings about how IE has evolved in recent years. Indeed, there is reason for both optimism and concern. While the overall incidence of IE remained stable during the study period, it is encouraging that 90-day mortality decreased, perhaps as a result of improvements in diagnosis and management, such as earlier valve replacement. There was a 38% increase in patients with IE who were hemodialysis-dependent, which could be interpreted as either hemodialysis patients have a higher risk for IE or perhaps there are more hemodialysis patients now in the general population. Although further investigation about the risks associated with hemodialysis and IE is needed, clinicians currently caring for hemodialysis patients need to be vigilant and maintain a high index of suspicion for IE, especially in the setting of a bloodstream infection.

Regarding the decrease in oral streptococcal IE, this trend could mean more people are receiving better dental care. That no increase in IE cases was observed after 2007 lends further support to the current IE prophylaxis recommendations. Drug use-associated IE increased over the study period, which serves as an important reminder about how the ongoing epidemic of intravenous drug abuse (especially heroin) is a major and very costly problem for society. Thus, one could argue from a purely economic standpoint that more resources (e.g., needle-exchange programs and funding for substance-abuse programs) should be devoted to this issue. The increase in MRSA IE is not surprising given the higher incidence of community-associated infections that began in the 1990s. What is unexpected given the high virulence of MRSA is that the overall mortality of IE declined during the study period. Perhaps the increase in MRSA cases was balanced by the decline of oral streptococcal ones.

As with all retrospective studies that use large databases, there is a chance that misclassification errors and unrecognized confounding variables affected the results. The IE data from New York and California might not be representative of other regions of the country, e.g., the Southern and Midwestern states. Moreover, the organisms were identified by ICD-9 codes and were assumed to be the causative pathogens, which may not have been the case. The data are already four years old and may not accurately reflect the current characteristics of IE. For example, there is evidence that the incidence of healthcare-associated MRSA infections is decreasing.1 Finally, the investigators were not able to identify cases of IE acquired in skilled care facilities, which could have led to an underestimate of healthcare-associated cases.

The study by Toyoda et al presents a lot of interesting data that can serve as a starting point for many future investigations. IE is an important and dynamic disease whose trends must be monitored continuously to achieve optimal patient outcomes.


  1. Evans ME, Kralovic SM, Simbartl LA, et al. Eight years of decreased methicillin-resistant Staphylococcus aureus health care-associated infections associated with a Veterans Affairs prevention initiative. Am J Infect Control 2017;45:13-16.